{ "retrieved_date": "2026-10-08", "scope": "Four requested families; all eight GSQ-RCO tiers; both BF16 checkpoints; all 27 publisher standard Swift GGUF tiers; ten Unsloth normal-Qwen quants; three linked Swift accelerator formats. Not an exhaustive catalog of every third-party conversion. MTP and vision projectors excluded as standalone model tiers.", "sources": { "1": { "url": "https://huggingface.co/Qwen/Qwen3.8-Flash-Next", "revision": "de4b8e4d43b917e7706784d8bb445c9af86a3540", "card_url": "https://huggingface.co/Qwen/Qwen3.8-Flash-Next/blob/de4b8e4d43b917e7706784d8bb445c9af86a3540/README.md" }, "2": { "url": "https://huggingface.co/ukisai/Swift1.5-Qwen3.8-Flash-Next", "revision": "0bd4fe22431372cdad1979267d3ab45aa7e6150a", "card_url": "https://huggingface.co/ukisai/Swift1.5-Qwen3.8-Flash-Next/blob/0bd4fe22431372cdad1979267d3ab45aa7e6150a/README.md" }, "3": { "url": "https://huggingface.co/ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF", "revision": "ed59f92082b1e93c0e96d60a8b11aab089b52f09", "card_url": "https://huggingface.co/ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF/blob/ed59f92082b1e93c0e96d60a8b11aab089b52f09/README.md" }, "4": { "url": "https://huggingface.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-GSQ-RCO-GGUF", "revision": "99bb8f7f95c7aa7b24a36a7786a4f657b30f5d3d", "card_url": "https://huggingface.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-GSQ-RCO-GGUF/blob/99bb8f7f95c7aa7b24a36a7786a4f657b30f5d3d/README.md" }, "5": { "url": "https://huggingface.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "revision": "0142554c963a8653912c0c8283f05f13f63653f3", "card_url": "https://huggingface.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF/blob/0142554c963a8653912c0c8283f05f13f63653f3/README.md" }, "6": { "url": "https://huggingface.co/unsloth/Qwen3.8-Flash-Next-GGUF", "revision": "766911a6b7369840a91dbcd95f9f997acaab6cd6", "card_url": "https://huggingface.co/unsloth/Qwen3.8-Flash-Next-GGUF/blob/766911a6b7369840a91dbcd95f9f997acaab6cd6/README.md" }, "10": { "url": "https://huggingface.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-W4A16-AWQ", "revision": "598afc96128f76a67117caf878f538d34f6f1e02", "card_url": "https://huggingface.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-W4A16-AWQ/blob/598afc96128f76a67117caf878f538d34f6f1e02/README.md" }, "11": { "url": "https://huggingface.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-W4A16-AutoRound", "revision": "177e8a382da63080bfd2875ab379fd8cb23e0dea", "card_url": "https://huggingface.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-W4A16-AutoRound/blob/177e8a382da63080bfd2875ab379fd8cb23e0dea/README.md" }, "12": { "url": "https://huggingface.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-NVFP4", "revision": "3ff0520224f264a2d0ac4ab56ece8f2f13aadb38", "card_url": "https://huggingface.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-NVFP4/blob/3ff0520224f264a2d0ac4ab56ece8f2f13aadb38/README.md" }, "7": { "url": "https://unsloth.ai/docs/models/qwen3.8-next", "retrieved_at": "2026-10-08T18:31:32.488342+00:00" }, "8": { "url": "https://github.com/Niko1221/Strata/blob/82f46a8c8f475f001ad76d92f58f4a4f8ffb0253/docs/MODELS.md" }, "9": { "url": "https://huggingface.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-GSQ-RCO-GGUF/blob/99bb8f7f95c7aa7b24a36a7786a4f657b30f5d3d/evaluation/heldout-kld.tsv" }, "13": { "url": "https://github.com/Niko1221/Strata/blob/82f46a8c8f475f001ad76d92f58f4a4f8ffb0253/docs/UNSLOTH_Q4.md" } }, "variants": [ { "group": "core", "family": "Normal Qwen", "tier": "BF16", "model_id": "Qwen/Qwen3.8-Flash-Next", "source": 1, "size_bytes": 360000192888, "files": [ { "rfilename": "model-00001-of-00131.safetensors", "size": 1040155944 }, { "rfilename": "model-00002-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00003-of-00131.safetensors", "size": 1832518264 }, { "rfilename": "model-00004-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00005-of-00131.safetensors", "size": 3355888504 }, { "rfilename": "model-00006-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00007-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00008-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00009-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00010-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00011-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00012-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00013-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00014-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00015-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00016-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00017-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00018-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00019-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00020-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00021-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00022-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00023-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00024-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00025-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00026-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00027-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00028-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00029-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00030-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00031-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "model-00032-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "model-00033-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "model-00034-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "model-00035-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "model-00036-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "model-00037-of-00131.safetensors", "size": 1742347160 }, { "rfilename": "model-00038-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00039-of-00131.safetensors", "size": 1716599840 }, { "rfilename": "model-00040-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00041-of-00131.safetensors", "size": 1941966288 }, { "rfilename": "model-00042-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00043-of-00131.safetensors", "size": 1839072016 }, { "rfilename": "model-00044-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00045-of-00131.safetensors", "size": 1832518280 }, { "rfilename": "model-00046-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00047-of-00131.safetensors", "size": 1716681904 }, { "rfilename": "model-00048-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00049-of-00131.safetensors", "size": 1935433152 }, { "rfilename": "model-00050-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00051-of-00131.safetensors", "size": 1832518280 }, { "rfilename": "model-00052-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00053-of-00131.safetensors", "size": 1839072024 }, { "rfilename": "model-00054-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00055-of-00131.safetensors", "size": 1716599840 }, { "rfilename": "model-00056-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00057-of-00131.safetensors", "size": 1941966280 }, { "rfilename": "model-00058-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00059-of-00131.safetensors", "size": 1716600048 }, { "rfilename": "model-00060-of-00131.safetensors", "size": 3471361528 }, { "rfilename": "model-00061-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00062-of-00131.safetensors", "size": 3510240000 }, { "rfilename": "model-00063-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00064-of-00131.safetensors", "size": 1832518280 }, { "rfilename": "model-00065-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00066-of-00131.safetensors", "size": 1719221392 }, { "rfilename": "model-00067-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00068-of-00131.safetensors", "size": 1935412552 }, { "rfilename": "model-00069-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00070-of-00131.safetensors", "size": 1832523520 }, { "rfilename": "model-00071-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00072-of-00131.safetensors", "size": 1839072128 }, { "rfilename": "model-00073-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00074-of-00131.safetensors", "size": 1716599840 }, { "rfilename": "model-00075-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00076-of-00131.safetensors", "size": 1938689480 }, { "rfilename": "model-00077-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00078-of-00131.safetensors", "size": 1832600344 }, { "rfilename": "model-00079-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00080-of-00131.safetensors", "size": 1716599840 }, { "rfilename": "model-00081-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00082-of-00131.safetensors", "size": 1935433136 }, { "rfilename": "model-00083-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00084-of-00131.safetensors", "size": 1716599840 }, { "rfilename": "model-00085-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00086-of-00131.safetensors", "size": 1941966296 }, { "rfilename": "model-00087-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00088-of-00131.safetensors", "size": 1839072024 }, { "rfilename": "model-00089-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00090-of-00131.safetensors", "size": 1835795216 }, { "rfilename": "model-00091-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00092-of-00131.safetensors", "size": 1716599840 }, { "rfilename": "model-00093-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00094-of-00131.safetensors", "size": 1935412928 }, { "rfilename": "model-00095-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00096-of-00131.safetensors", "size": 1832518280 }, { "rfilename": "model-00097-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00098-of-00131.safetensors", "size": 1832518280 }, { "rfilename": "model-00099-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00100-of-00131.safetensors", "size": 1716600224 }, { "rfilename": "model-00101-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00102-of-00131.safetensors", "size": 1935413160 }, { "rfilename": "model-00103-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00104-of-00131.safetensors", "size": 1832518272 }, { "rfilename": "model-00105-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00106-of-00131.safetensors", "size": 1863975680 }, { "rfilename": "model-00107-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00108-of-00131.safetensors", "size": 1900676064 }, { "rfilename": "model-00109-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00110-of-00131.safetensors", "size": 1716599840 }, { "rfilename": "model-00111-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00112-of-00131.safetensors", "size": 1935413176 }, { "rfilename": "model-00113-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00114-of-00131.safetensors", "size": 1832523512 }, { "rfilename": "model-00115-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00116-of-00131.safetensors", "size": 1832518280 }, { "rfilename": "model-00117-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00118-of-00131.safetensors", "size": 1716620424 }, { "rfilename": "model-00119-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00120-of-00131.safetensors", "size": 1948519840 }, { "rfilename": "model-00121-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00122-of-00131.safetensors", "size": 1832518264 }, { "rfilename": "model-00123-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00124-of-00131.safetensors", "size": 1729707128 }, { "rfilename": "model-00125-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00126-of-00131.safetensors", "size": 1935412520 }, { "rfilename": "model-00127-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00128-of-00131.safetensors", "size": 1832518264 }, { "rfilename": "model-00129-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00130-of-00131.safetensors", "size": 2974788136 }, { "rfilename": "model-00131-of-00131.safetensors", "size": 1271398528 } ], "hardware": "~360 GB checkpoint; multi-GPU/server-class capacity plus runtime headroom. Not a Strata installer option.", "kld": null, "quality": "BF16 reference; see model-level benchmarks.", "decode": null }, { "group": "core", "family": "Swift 1.5 normal", "tier": "BF16", "model_id": "ukisai/Swift1.5-Qwen3.8-Flash-Next", "source": 2, "size_bytes": 360000192888, "files": [ { "rfilename": "model-00001-of-00131.safetensors", "size": 1040155944 }, { "rfilename": "model-00002-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00003-of-00131.safetensors", "size": 1832518264 }, { "rfilename": "model-00004-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00005-of-00131.safetensors", "size": 3355888504 }, { "rfilename": "model-00006-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00007-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00008-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00009-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00010-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00011-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00012-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00013-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00014-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00015-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00016-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00017-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00018-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00019-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00020-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00021-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00022-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00023-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00024-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00025-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00026-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00027-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00028-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00029-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00030-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "model-00031-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "model-00032-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "model-00033-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "model-00034-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "model-00035-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "model-00036-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "model-00037-of-00131.safetensors", "size": 1742347160 }, { "rfilename": "model-00038-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00039-of-00131.safetensors", "size": 1716599840 }, { "rfilename": "model-00040-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00041-of-00131.safetensors", "size": 1941966288 }, { "rfilename": "model-00042-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00043-of-00131.safetensors", "size": 1839072016 }, { "rfilename": "model-00044-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00045-of-00131.safetensors", "size": 1832518280 }, { "rfilename": "model-00046-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00047-of-00131.safetensors", "size": 1716681904 }, { "rfilename": "model-00048-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00049-of-00131.safetensors", "size": 1935433152 }, { "rfilename": "model-00050-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00051-of-00131.safetensors", "size": 1832518280 }, { "rfilename": "model-00052-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00053-of-00131.safetensors", "size": 1839072024 }, { "rfilename": "model-00054-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00055-of-00131.safetensors", "size": 1716599840 }, { "rfilename": "model-00056-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00057-of-00131.safetensors", "size": 1941966280 }, { "rfilename": "model-00058-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00059-of-00131.safetensors", "size": 1716600048 }, { "rfilename": "model-00060-of-00131.safetensors", "size": 3471361528 }, { "rfilename": "model-00061-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00062-of-00131.safetensors", "size": 3510240000 }, { "rfilename": "model-00063-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00064-of-00131.safetensors", "size": 1832518280 }, { "rfilename": "model-00065-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00066-of-00131.safetensors", "size": 1719221392 }, { "rfilename": "model-00067-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00068-of-00131.safetensors", "size": 1935412552 }, { "rfilename": "model-00069-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00070-of-00131.safetensors", "size": 1832523520 }, { "rfilename": "model-00071-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00072-of-00131.safetensors", "size": 1839072128 }, { "rfilename": "model-00073-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00074-of-00131.safetensors", "size": 1716599840 }, { "rfilename": "model-00075-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00076-of-00131.safetensors", "size": 1938689480 }, { "rfilename": "model-00077-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00078-of-00131.safetensors", "size": 1832600344 }, { "rfilename": "model-00079-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00080-of-00131.safetensors", "size": 1716599840 }, { "rfilename": "model-00081-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00082-of-00131.safetensors", "size": 1935433136 }, { "rfilename": "model-00083-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00084-of-00131.safetensors", "size": 1716599840 }, { "rfilename": "model-00085-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00086-of-00131.safetensors", "size": 1941966296 }, { "rfilename": "model-00087-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00088-of-00131.safetensors", "size": 1839072024 }, { "rfilename": "model-00089-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00090-of-00131.safetensors", "size": 1835795216 }, { "rfilename": "model-00091-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00092-of-00131.safetensors", "size": 1716599840 }, { "rfilename": "model-00093-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00094-of-00131.safetensors", "size": 1935412928 }, { "rfilename": "model-00095-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00096-of-00131.safetensors", "size": 1832518280 }, { "rfilename": "model-00097-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00098-of-00131.safetensors", "size": 1832518280 }, { "rfilename": "model-00099-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00100-of-00131.safetensors", "size": 1716600224 }, { "rfilename": "model-00101-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00102-of-00131.safetensors", "size": 1935413160 }, { "rfilename": "model-00103-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00104-of-00131.safetensors", "size": 1832518272 }, { "rfilename": "model-00105-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00106-of-00131.safetensors", "size": 1863975680 }, { "rfilename": "model-00107-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00108-of-00131.safetensors", "size": 1900676064 }, { "rfilename": "model-00109-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00110-of-00131.safetensors", "size": 1716599840 }, { "rfilename": "model-00111-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00112-of-00131.safetensors", "size": 1935413176 }, { "rfilename": "model-00113-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00114-of-00131.safetensors", "size": 1832523512 }, { "rfilename": "model-00115-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00116-of-00131.safetensors", "size": 1832518280 }, { "rfilename": "model-00117-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00118-of-00131.safetensors", "size": 1716620424 }, { "rfilename": "model-00119-of-00131.safetensors", "size": 3355443376 }, { "rfilename": "model-00120-of-00131.safetensors", "size": 1948519840 }, { "rfilename": "model-00121-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00122-of-00131.safetensors", "size": 1832518264 }, { "rfilename": "model-00123-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00124-of-00131.safetensors", "size": 1729707128 }, { "rfilename": "model-00125-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00126-of-00131.safetensors", "size": 1935412520 }, { "rfilename": "model-00127-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00128-of-00131.safetensors", "size": 1832518264 }, { "rfilename": "model-00129-of-00131.safetensors", "size": 3355443368 }, { "rfilename": "model-00130-of-00131.safetensors", "size": 2974788136 }, { "rfilename": "model-00131-of-00131.safetensors", "size": 1271398528 } ], "hardware": "~360 GB checkpoint; multi-GPU/server-class capacity plus runtime headroom. Not a Strata installer option.", "kld": null, "quality": "BF16 reference; see model-level benchmarks.", "decode": null }, { "group": "core", "family": "Qwen GSQ-RCO", "tier": "Q2_0", "model_id": "ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF", "source": 3, "size_bytes": 66423878624, "files": [ { "rfilename": "Q2_0/Qwen3.8-Flash-Next-GSQ-RCO-Q2_0-00001-of-00002.gguf", "size": 37623740192 }, { "rfilename": "Q2_0/Qwen3.8-Flash-Next-GSQ-RCO-Q2_0-00002-of-00002.gguf", "size": 28800138432 } ], "hardware": "Planning: 48 GB RAM + 12 GB+ supported GPU + SSD; cache/context/OS need headroom.", "kld": { "value": 0.218335, "context": 512, "corpus": "English C4 prose, 100 chunks", "reference": "Qwen BF16", "source": 9 }, "quality": { "AIME25": 96.67, "GPQA-D": 89.39, "LCB v6": 81.14, "Task average": 89.07 }, "decode": { "tps": 94, "source": 8, "prefill_tps": 2650, "long_context_tps": 76, "hardware": "RTX 5070 12 GB; Ryzen 5 7600; 64 GB RAM; Strata.", "notes": "4K answer / short chat; prefill measured at 32K prompt; Q2_0 engine 0.1.36, other tiers 0.1.26." } }, { "group": "core", "family": "Qwen GSQ-RCO", "tier": "IQ2_XS", "model_id": "ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF", "source": 3, "size_bytes": 68026093024, "files": [ { "rfilename": "IQ2_XS/Qwen3.8-Flash-Next-GSQ-RCO-IQ2_XS-00001-of-00002.gguf", "size": 39225954592 }, { "rfilename": "IQ2_XS/Qwen3.8-Flash-Next-GSQ-RCO-IQ2_XS-00002-of-00002.gguf", "size": 28800138432 } ], "hardware": "Planning: 48 GB RAM + 12 GB+ supported GPU + SSD; cache/context/OS need headroom.", "kld": { "value": 0.200267, "context": 512, "corpus": "English C4 prose, 100 chunks", "reference": "Qwen BF16", "source": 9 }, "quality": { "AIME25": 96.67, "GPQA-D": 87.37, "LCB v6": 83.43, "Task average": 89.16 }, "decode": { "tps": 79, "source": 8, "prefill_tps": 2090, "long_context_tps": 63, "hardware": "RTX 5070 12 GB; Ryzen 5 7600; 64 GB RAM; Strata.", "notes": "4K answer / short chat; prefill measured at 32K prompt; Q2_0 engine 0.1.36, other tiers 0.1.26." } }, { "group": "core", "family": "Qwen GSQ-RCO", "tier": "IQ3_XXS", "model_id": "ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF", "source": 3, "size_bytes": 75839998528, "files": [ { "rfilename": "IQ3_XXS/Qwen3.8-Flash-Next-GSQ-RCO-IQ3_XXS-00001-of-00002.gguf", "size": 47039860096 }, { "rfilename": "IQ3_XXS/Qwen3.8-Flash-Next-GSQ-RCO-IQ3_XXS-00002-of-00002.gguf", "size": 28800138432 } ], "hardware": "Planning: 64 GB RAM + 12 GB+ supported GPU + SSD; cache/context/OS need headroom.", "kld": { "value": 0.114982, "context": 512, "corpus": "English C4 prose, 100 chunks", "reference": "Qwen BF16", "source": 9 }, "quality": { "AIME25": 100.0, "GPQA-D": 91.41, "LCB v6": 86.29, "Task average": 92.57 }, "decode": { "tps": 62, "source": 8, "prefill_tps": 1750, "long_context_tps": 49, "hardware": "RTX 5070 12 GB; Ryzen 5 7600; 64 GB RAM; Strata.", "notes": "4K answer / short chat; prefill measured at 32K prompt; Q2_0 engine 0.1.36, other tiers 0.1.26." } }, { "group": "core", "family": "Qwen GSQ-RCO", "tier": "IQ3_S", "model_id": "ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF", "source": 3, "size_bytes": 83617662656, "files": [ { "rfilename": "IQ3_S/Qwen3.8-Flash-Next-GSQ-RCO-IQ3_S-00001-of-00002.gguf", "size": 54817524224 }, { "rfilename": "IQ3_S/Qwen3.8-Flash-Next-GSQ-RCO-IQ3_S-00002-of-00002.gguf", "size": 28800138432 } ], "hardware": "Planning: 64 GB RAM + 12 GB+ supported GPU + SSD; cache/context/OS need headroom.", "kld": { "value": 0.062341, "context": 512, "corpus": "English C4 prose, 100 chunks", "reference": "Qwen BF16", "source": 9 }, "quality": { "AIME25": 100.0, "GPQA-D": 92.93, "LCB v6": 86.86, "Task average": 93.26 }, "decode": { "tps": 53, "source": 8, "prefill_tps": 1620, "long_context_tps": 46, "hardware": "RTX 5070 12 GB; Ryzen 5 7600; 64 GB RAM; Strata.", "notes": "4K answer / short chat; prefill measured at 32K prompt; Q2_0 engine 0.1.36, other tiers 0.1.26." } }, { "group": "core", "family": "Swift 1.5 GSQ-RCO", "tier": "Q2_0", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GSQ-RCO-GGUF", "source": 4, "size_bytes": 66549952800, "files": [ { "rfilename": "Swift-Qwen3.8-Flash-Next-GSQ-RCO-Q2_0-00001-of-00002.gguf", "size": 39799117984 }, { "rfilename": "Swift-Qwen3.8-Flash-Next-GSQ-RCO-Q2_0-00002-of-00002.gguf", "size": 26750834816 } ], "hardware": "Planning: 48 GB RAM + 12 GB+ supported GPU + SSD; cache/context/OS need headroom. Same-tier estimate; current Swift release not fit-tested here.", "kld": { "value": 0.234242, "context": 512, "corpus": "English C4 prose, 100 chunks", "reference": "Swift BF16", "source": 9 }, "quality": "Quant task accuracy not reported in this source; KLD is not task accuracy.", "decode": null }, { "group": "core", "family": "Swift 1.5 GSQ-RCO", "tier": "IQ2_XS", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GSQ-RCO-GGUF", "source": 4, "size_bytes": 68152167168, "files": [ { "rfilename": "Swift-Qwen3.8-Flash-Next-GSQ-RCO-IQ2_XS-00001-of-00002.gguf", "size": 39788473344 }, { "rfilename": "Swift-Qwen3.8-Flash-Next-GSQ-RCO-IQ2_XS-00002-of-00002.gguf", "size": 28363693824 } ], "hardware": "Planning: 48 GB RAM + 12 GB+ supported GPU + SSD; cache/context/OS need headroom. Same-tier estimate; current Swift release not fit-tested here.", "kld": { "value": 0.188117, "context": 512, "corpus": "English C4 prose, 100 chunks", "reference": "Swift BF16", "source": 9 }, "quality": "Quant task accuracy not reported in this source; KLD is not task accuracy.", "decode": { "tps": 126.0, "source": "local", "notes": "One installed-model run, 265 generated tokens including thinking; not a sweep or a proof for the current HF release.", "hardware": "RTX 5090 Laptop 24 GB; Core Ultra 9 275HX; 64 GB RAM; Strata; configured max context 262144, actual prompt length unrecorded." } }, { "group": "core", "family": "Swift 1.5 GSQ-RCO", "tier": "IQ3_XXS", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GSQ-RCO-GGUF", "source": 4, "size_bytes": 75966073120, "files": [ { "rfilename": "Swift-Qwen3.8-Flash-Next-GSQ-RCO-IQ3_XXS-00001-of-00002.gguf", "size": 39785790560 }, { "rfilename": "Swift-Qwen3.8-Flash-Next-GSQ-RCO-IQ3_XXS-00002-of-00002.gguf", "size": 36180282560 } ], "hardware": "Planning: 64 GB RAM + 12 GB+ supported GPU + SSD; cache/context/OS need headroom. Same-tier estimate; current Swift release not fit-tested here.", "kld": { "value": 0.116077, "context": 512, "corpus": "English C4 prose, 100 chunks", "reference": "Swift BF16", "source": 9 }, "quality": "Quant task accuracy not reported in this source; KLD is not task accuracy.", "decode": null }, { "group": "core", "family": "Swift 1.5 GSQ-RCO", "tier": "IQ3_S", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GSQ-RCO-GGUF", "source": 4, "size_bytes": 83743737120, "files": [ { "rfilename": "Swift-Qwen3.8-Flash-Next-GSQ-RCO-IQ3_S-00001-of-00002.gguf", "size": 44900816736 }, { "rfilename": "Swift-Qwen3.8-Flash-Next-GSQ-RCO-IQ3_S-00002-of-00002.gguf", "size": 38842920384 } ], "hardware": "Planning: 64 GB RAM + 12 GB+ supported GPU + SSD; cache/context/OS need headroom. Same-tier estimate; current Swift release not fit-tested here.", "kld": { "value": 0.062556, "context": 512, "corpus": "English C4 prose, 100 chunks", "reference": "Swift BF16", "source": 9 }, "quality": "Quant task accuracy not reported in this source; KLD is not task accuracy.", "decode": null }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q8_0", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 188260769696, "files": [ { "rfilename": "Q8_0/Swift-1.5-Qwen3.8-Flash-Next-Q8_0-00001-of-00005.gguf", "size": 699524416 }, { "rfilename": "Q8_0/Swift-1.5-Qwen3.8-Flash-Next-Q8_0-00002-of-00005.gguf", "size": 54400261312 }, { "rfilename": "Q8_0/Swift-1.5-Qwen3.8-Flash-Next-Q8_0-00003-of-00005.gguf", "size": 44905193504 }, { "rfilename": "Q8_0/Swift-1.5-Qwen3.8-Flash-Next-Q8_0-00004-of-00005.gguf", "size": 44150177536 }, { "rfilename": "Q8_0/Swift-1.5-Qwen3.8-Flash-Next-Q8_0-00005-of-00005.gguf", "size": 44105612928 } ], "hardware": "Conservative full-file-resident plan (estimate): ~224 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.0297, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.022, "long_context": 32768, "top_token_agreement": "94.91%" }, "quality": "Top-token agreement at 32K: 94.91%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 224, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q6_K", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 168113805248, "files": [ { "rfilename": "Q6_K/Swift-1.5-Qwen3.8-Flash-Next-Q6_K-00001-of-00005.gguf", "size": 545566016 }, { "rfilename": "Q6_K/Swift-1.5-Qwen3.8-Flash-Next-Q6_K-00002-of-00005.gguf", "size": 54400261312 }, { "rfilename": "Q6_K/Swift-1.5-Qwen3.8-Flash-Next-Q6_K-00003-of-00005.gguf", "size": 44411907360 }, { "rfilename": "Q6_K/Swift-1.5-Qwen3.8-Flash-Next-Q6_K-00004-of-00005.gguf", "size": 44558394624 }, { "rfilename": "Q6_K/Swift-1.5-Qwen3.8-Flash-Next-Q6_K-00005-of-00005.gguf", "size": 24197675936 } ], "hardware": "Conservative full-file-resident plan (estimate): ~192 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.0324, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.0252, "long_context": 32768, "top_token_agreement": "94.34%" }, "quality": "Top-token agreement at 32K: 94.34%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 192, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q5_K_L", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 151061061568, "files": [ { "rfilename": "Q5_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q5_K_L-00001-of-00005.gguf", "size": 699524416 }, { "rfilename": "Q5_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q5_K_L-00002-of-00005.gguf", "size": 54400261312 }, { "rfilename": "Q5_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q5_K_L-00003-of-00005.gguf", "size": 44399498048 }, { "rfilename": "Q5_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q5_K_L-00004-of-00005.gguf", "size": 44603368576 }, { "rfilename": "Q5_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q5_K_L-00005-of-00005.gguf", "size": 6958409216 } ], "hardware": "Conservative full-file-resident plan (estimate): ~192 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.0788, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.0575, "long_context": 32768, "top_token_agreement": "91.61%" }, "quality": "Top-token agreement at 32K: 91.61%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 192, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q4_K_L", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 139270930272, "files": [ { "rfilename": "Q4_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q4_K_L-00001-of-00004.gguf", "size": 699524416 }, { "rfilename": "Q4_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q4_K_L-00002-of-00004.gguf", "size": 54400261312 }, { "rfilename": "Q4_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q4_K_L-00003-of-00004.gguf", "size": 44612208064 }, { "rfilename": "Q4_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q4_K_L-00004-of-00004.gguf", "size": 39558936480 } ], "hardware": "Conservative full-file-resident plan (estimate): ~160 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.1051, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.074, "long_context": 32768, "top_token_agreement": "90.14%" }, "quality": "Top-token agreement at 32K: 90.14%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 160, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q5_K_M", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 134668639040, "files": [ { "rfilename": "Q5_K_M/Swift-1.5-Qwen3.8-Flash-Next-Q5_K_M-00001-of-00004.gguf", "size": 44580008800 }, { "rfilename": "Q5_K_M/Swift-1.5-Qwen3.8-Flash-Next-Q5_K_M-00002-of-00004.gguf", "size": 44975799264 }, { "rfilename": "Q5_K_M/Swift-1.5-Qwen3.8-Flash-Next-Q5_K_M-00003-of-00004.gguf", "size": 44521427968 }, { "rfilename": "Q5_K_M/Swift-1.5-Qwen3.8-Flash-Next-Q5_K_M-00004-of-00004.gguf", "size": 591403008 } ], "hardware": "Conservative full-file-resident plan (estimate): ~160 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.1641, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.1386, "long_context": 32768, "top_token_agreement": "87.34%" }, "quality": "Top-token agreement at 32K: 87.34%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 160, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q5_K_S", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 128104512224, "files": [ { "rfilename": "Q5_K_S/Swift-1.5-Qwen3.8-Flash-Next-Q5_K_S-00001-of-00003.gguf", "size": 45003631968 }, { "rfilename": "Q5_K_S/Swift-1.5-Qwen3.8-Flash-Next-Q5_K_S-00002-of-00003.gguf", "size": 44956919360 }, { "rfilename": "Q5_K_S/Swift-1.5-Qwen3.8-Flash-Next-Q5_K_S-00003-of-00003.gguf", "size": 38143960896 } ], "hardware": "Conservative full-file-resident plan (estimate): ~160 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.1777, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.1482, "long_context": 32768, "top_token_agreement": "86.92%" }, "quality": "Top-token agreement at 32K: 86.92%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 160, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q4_K_M", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 119599029952, "files": [ { "rfilename": "Q4_K_M/Swift-1.5-Qwen3.8-Flash-Next-Q4_K_M-00001-of-00003.gguf", "size": 44648342368 }, { "rfilename": "Q4_K_M/Swift-1.5-Qwen3.8-Flash-Next-Q4_K_M-00002-of-00003.gguf", "size": 44928589600 }, { "rfilename": "Q4_K_M/Swift-1.5-Qwen3.8-Flash-Next-Q4_K_M-00003-of-00003.gguf", "size": 30022097984 } ], "hardware": "Conservative full-file-resident plan (estimate): ~160 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.1512, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.11, "long_context": 32768, "top_token_agreement": "88.20%" }, "quality": "Top-token agreement at 32K: 88.20%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 160, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q3_K_XL", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 119347361600, "files": [ { "rfilename": "Q3_K_XL/Swift-1.5-Qwen3.8-Flash-Next-Q3_K_XL-00001-of-00004.gguf", "size": 699524416 }, { "rfilename": "Q3_K_XL/Swift-1.5-Qwen3.8-Flash-Next-Q3_K_XL-00002-of-00004.gguf", "size": 54400261312 }, { "rfilename": "Q3_K_XL/Swift-1.5-Qwen3.8-Flash-Next-Q3_K_XL-00003-of-00004.gguf", "size": 44985977344 }, { "rfilename": "Q3_K_XL/Swift-1.5-Qwen3.8-Flash-Next-Q3_K_XL-00004-of-00004.gguf", "size": 19261598528 } ], "hardware": "Conservative full-file-resident plan (estimate): ~160 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.3718, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.325, "long_context": 32768, "top_token_agreement": "79.79%" }, "quality": "Top-token agreement at 32K: 79.79%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 160, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q4_K_S", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 113135910592, "files": [ { "rfilename": "Q4_K_S/Swift-1.5-Qwen3.8-Flash-Next-Q4_K_S-00001-of-00003.gguf", "size": 44958237952 }, { "rfilename": "Q4_K_S/Swift-1.5-Qwen3.8-Flash-Next-Q4_K_S-00002-of-00003.gguf", "size": 44615499136 }, { "rfilename": "Q4_K_S/Swift-1.5-Qwen3.8-Flash-Next-Q4_K_S-00003-of-00003.gguf", "size": 23562173504 } ], "hardware": "Conservative full-file-resident plan (estimate): ~160 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.1779, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.1311, "long_context": 32768, "top_token_agreement": "87.01%" }, "quality": "Top-token agreement at 32K: 87.01%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 160, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q4_1", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 111258575552, "files": [ { "rfilename": "Q4_1/Swift-1.5-Qwen3.8-Flash-Next-Q4_1-00001-of-00003.gguf", "size": 44954809632 }, { "rfilename": "Q4_1/Swift-1.5-Qwen3.8-Flash-Next-Q4_1-00002-of-00003.gguf", "size": 44528290848 }, { "rfilename": "Q4_1/Swift-1.5-Qwen3.8-Flash-Next-Q4_1-00003-of-00003.gguf", "size": 21775475072 } ], "hardware": "Conservative full-file-resident plan (estimate): ~128 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.2532, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.1812, "long_context": 32768, "top_token_agreement": "85.07%" }, "quality": "Top-token agreement at 32K: 85.07%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 128, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q2_K_L", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 107147702080, "files": [ { "rfilename": "Q2_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q2_K_L-00001-of-00004.gguf", "size": 699524416 }, { "rfilename": "Q2_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q2_K_L-00002-of-00004.gguf", "size": 54400261312 }, { "rfilename": "Q2_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q2_K_L-00003-of-00004.gguf", "size": 44998709312 }, { "rfilename": "Q2_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q2_K_L-00004-of-00004.gguf", "size": 7049207040 } ], "hardware": "Conservative full-file-resident plan (estimate): ~128 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.5425, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.3867, "long_context": 32768, "top_token_agreement": "77.28%" }, "quality": "Top-token agreement at 32K: 77.28%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 128, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q4_0", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 100599252672, "files": [ { "rfilename": "Q4_0/Swift-1.5-Qwen3.8-Flash-Next-Q4_0-00001-of-00003.gguf", "size": 44772642144 }, { "rfilename": "Q4_0/Swift-1.5-Qwen3.8-Flash-Next-Q4_0-00002-of-00003.gguf", "size": 44997356288 }, { "rfilename": "Q4_0/Swift-1.5-Qwen3.8-Flash-Next-Q4_0-00003-of-00003.gguf", "size": 10829254240 } ], "hardware": "Conservative full-file-resident plan (estimate): ~128 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.2524, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.2593, "long_context": 32768, "top_token_agreement": "82.53%" }, "quality": "Top-token agreement at 32K: 82.53%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 128, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "IQ4_NL", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 100293527232, "files": [ { "rfilename": "IQ4_NL/Swift-1.5-Qwen3.8-Flash-Next-IQ4_NL-00001-of-00003.gguf", "size": 44932939232 }, { "rfilename": "IQ4_NL/Swift-1.5-Qwen3.8-Flash-Next-IQ4_NL-00002-of-00003.gguf", "size": 44533986112 }, { "rfilename": "IQ4_NL/Swift-1.5-Qwen3.8-Flash-Next-IQ4_NL-00003-of-00003.gguf", "size": 10826601888 } ], "hardware": "Conservative full-file-resident plan (estimate): ~128 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.1261, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.0926, "long_context": 32768, "top_token_agreement": "88.96%" }, "quality": "Top-token agreement at 32K: 88.96%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 128, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "IQ4_XS", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 97681835744, "files": [ { "rfilename": "IQ4_XS/Swift-1.5-Qwen3.8-Flash-Next-IQ4_XS-00001-of-00003.gguf", "size": 44822920960 }, { "rfilename": "IQ4_XS/Swift-1.5-Qwen3.8-Flash-Next-IQ4_XS-00002-of-00003.gguf", "size": 44822625408 }, { "rfilename": "IQ4_XS/Swift-1.5-Qwen3.8-Flash-Next-IQ4_XS-00003-of-00003.gguf", "size": 8036289376 } ], "hardware": "Conservative full-file-resident plan (estimate): ~128 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.1305, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.0929, "long_context": 32768, "top_token_agreement": "88.91%" }, "quality": "Top-token agreement at 32K: 88.91%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 128, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "IQ3_M", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 93227374272, "files": [ { "rfilename": "IQ3_M/Swift-1.5-Qwen3.8-Flash-Next-IQ3_M-00001-of-00003.gguf", "size": 44479066112 }, { "rfilename": "IQ3_M/Swift-1.5-Qwen3.8-Flash-Next-IQ3_M-00002-of-00003.gguf", "size": 44543324384 }, { "rfilename": "IQ3_M/Swift-1.5-Qwen3.8-Flash-Next-IQ3_M-00003-of-00003.gguf", "size": 4204983776 } ], "hardware": "Conservative full-file-resident plan (estimate): ~128 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.2005, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.1482, "long_context": 32768, "top_token_agreement": "86.22%" }, "quality": "Top-token agreement at 32K: 86.22%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 128, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q3_K_L", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 93191001792, "files": [ { "rfilename": "Q3_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q3_K_L-00001-of-00003.gguf", "size": 44468948992 }, { "rfilename": "Q3_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q3_K_L-00002-of-00003.gguf", "size": 44521124064 }, { "rfilename": "Q3_K_L/Swift-1.5-Qwen3.8-Flash-Next-Q3_K_L-00003-of-00003.gguf", "size": 4200928736 } ], "hardware": "Conservative full-file-resident plan (estimate): ~128 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.3842, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.33, "long_context": 32768, "top_token_agreement": "79.44%" }, "quality": "Top-token agreement at 32K: 79.44%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 128, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q3_K_M", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 91957778144, "files": [ { "rfilename": "Q3_K_M/Swift-1.5-Qwen3.8-Flash-Next-Q3_K_M-00001-of-00003.gguf", "size": 44686672000 }, { "rfilename": "Q3_K_M/Swift-1.5-Qwen3.8-Flash-Next-Q3_K_M-00002-of-00003.gguf", "size": 44582853152 }, { "rfilename": "Q3_K_M/Swift-1.5-Qwen3.8-Flash-Next-Q3_K_M-00003-of-00003.gguf", "size": 2688252992 } ], "hardware": "Conservative full-file-resident plan (estimate): ~128 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.321, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.226, "long_context": 32768, "top_token_agreement": "82.73%" }, "quality": "Top-token agreement at 32K: 82.73%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 128, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "IQ3_XS", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 91945490144, "files": [ { "rfilename": "IQ3_XS/Swift-1.5-Qwen3.8-Flash-Next-IQ3_XS-00001-of-00003.gguf", "size": 44684624000 }, { "rfilename": "IQ3_XS/Swift-1.5-Qwen3.8-Flash-Next-IQ3_XS-00002-of-00003.gguf", "size": 44572613152 }, { "rfilename": "IQ3_XS/Swift-1.5-Qwen3.8-Flash-Next-IQ3_XS-00003-of-00003.gguf", "size": 2688252992 } ], "hardware": "Conservative full-file-resident plan (estimate): ~128 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.2211, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.1644, "long_context": 32768, "top_token_agreement": "85.26%" }, "quality": "Top-token agreement at 32K: 85.26%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 128, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q3_K_S", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 89441195616, "files": [ { "rfilename": "Q3_K_S/Swift-1.5-Qwen3.8-Flash-Next-Q3_K_S-00001-of-00002.gguf", "size": 44612250016 }, { "rfilename": "Q3_K_S/Swift-1.5-Qwen3.8-Flash-Next-Q3_K_S-00002-of-00002.gguf", "size": 44828945600 } ], "hardware": "Conservative full-file-resident plan (estimate): ~128 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.3253, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.2284, "long_context": 32768, "top_token_agreement": "82.59%" }, "quality": "Top-token agreement at 32K: 82.59%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 128, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "IQ3_XXS", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 88021245536, "files": [ { "rfilename": "IQ3_XXS/Swift-1.5-Qwen3.8-Flash-Next-IQ3_XXS-00001-of-00002.gguf", "size": 44875407904 }, { "rfilename": "IQ3_XXS/Swift-1.5-Qwen3.8-Flash-Next-IQ3_XXS-00002-of-00002.gguf", "size": 43145837632 } ], "hardware": "Conservative full-file-resident plan (estimate): ~128 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.2868, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.2185, "long_context": 32768, "top_token_agreement": "83.30%" }, "quality": "Top-token agreement at 32K: 83.30%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 128, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "Q2_K", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 80926778976, "files": [ { "rfilename": "Q2_K/Swift-1.5-Qwen3.8-Flash-Next-Q2_K-00001-of-00002.gguf", "size": 44565396736 }, { "rfilename": "Q2_K/Swift-1.5-Qwen3.8-Flash-Next-Q2_K-00002-of-00002.gguf", "size": 36361382240 } ], "hardware": "Conservative full-file-resident plan (estimate): ~128 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.5703, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.4091, "long_context": 32768, "top_token_agreement": "76.66%" }, "quality": "Top-token agreement at 32K: 76.66%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 128, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "IQ2_M", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 80369487424, "files": [ { "rfilename": "IQ2_M/Swift-1.5-Qwen3.8-Flash-Next-IQ2_M-00001-of-00002.gguf", "size": 44868562304 }, { "rfilename": "IQ2_M/Swift-1.5-Qwen3.8-Flash-Next-IQ2_M-00002-of-00002.gguf", "size": 35500925120 } ], "hardware": "Conservative full-file-resident plan (estimate): ~128 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.3769, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.2759, "long_context": 32768, "top_token_agreement": "80.76%" }, "quality": "Top-token agreement at 32K: 80.76%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 128, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "IQ2_S", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 77825338976, "files": [ { "rfilename": "IQ2_S/Swift-1.5-Qwen3.8-Flash-Next-IQ2_S-00001-of-00002.gguf", "size": 44663336864 }, { "rfilename": "IQ2_S/Swift-1.5-Qwen3.8-Flash-Next-IQ2_S-00002-of-00002.gguf", "size": 33162002112 } ], "hardware": "Conservative full-file-resident plan (estimate): ~96 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.4846, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.4331, "long_context": 32768, "top_token_agreement": "76.47%" }, "quality": "Top-token agreement at 32K: 76.47%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 96, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "IQ2_XS", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 77737182816, "files": [ { "rfilename": "IQ2_XS/Swift-1.5-Qwen3.8-Flash-Next-IQ2_XS-00001-of-00002.gguf", "size": 44590909344 }, { "rfilename": "IQ2_XS/Swift-1.5-Qwen3.8-Flash-Next-IQ2_XS-00002-of-00002.gguf", "size": 33146273472 } ], "hardware": "Conservative full-file-resident plan (estimate): ~96 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.5136, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.4033, "long_context": 32768, "top_token_agreement": "76.93%" }, "quality": "Top-token agreement at 32K: 76.93%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 96, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "IQ2_XXS", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 75193034336, "files": [ { "rfilename": "IQ2_XXS/Swift-1.5-Qwen3.8-Flash-Next-IQ2_XXS-00001-of-00002.gguf", "size": 44744501568 }, { "rfilename": "IQ2_XXS/Swift-1.5-Qwen3.8-Flash-Next-IQ2_XXS-00002-of-00002.gguf", "size": 30448532768 } ], "hardware": "Conservative full-file-resident plan (estimate): ~96 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.6052, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.4472, "long_context": 32768, "top_token_agreement": "75.58%" }, "quality": "Top-token agreement at 32K: 75.58%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 96, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "IQ1_M", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 72012848704, "files": [ { "rfilename": "IQ1_M/Swift-1.5-Qwen3.8-Flash-Next-IQ1_M-00001-of-00002.gguf", "size": 44566676192 }, { "rfilename": "IQ1_M/Swift-1.5-Qwen3.8-Flash-Next-IQ1_M-00002-of-00002.gguf", "size": 27446172512 } ], "hardware": "Conservative full-file-resident plan (estimate): ~96 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 0.9067, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.6928, "long_context": 32768, "top_token_agreement": "69.57%" }, "quality": "Top-token agreement at 32K: 69.57%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 96, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "swift-standard", "family": "Swift 1.5 standard GGUF", "tier": "IQ1_S", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF", "source": 5, "size_bytes": 70104737376, "files": [ { "rfilename": "IQ1_S/Swift-1.5-Qwen3.8-Flash-Next-IQ1_S-00001-of-00002.gguf", "size": 44730559648 }, { "rfilename": "IQ1_S/Swift-1.5-Qwen3.8-Flash-Next-IQ1_S-00002-of-00002.gguf", "size": 25374177728 } ], "hardware": "Conservative full-file-resident plan (estimate): ~96 GB total RAM+VRAM or unified memory. Assumes weight bytes + 16 GB reserve, rounded to 32 GB; not a tested minimum. Sparse paging may use less; large contexts may need more.", "kld": { "value": 1.1601, "context": 512, "corpus": "WikiText-2 test, 100 windows", "reference": "Swift BF16", "source": 5, "long_context_value": 0.9343, "long_context": 32768, "top_token_agreement": "65.15%" }, "quality": "Top-token agreement at 32K: 65.15%; not task accuracy.", "decode": null, "memory_plan": { "total_gb": 96, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "normal-community", "family": "Normal Qwen / Unsloth GGUF", "tier": "UD-IQ1_S", "model_id": "unsloth/Qwen3.8-Flash-Next-GGUF", "source": 6, "size_bytes": 72546461344, "files": [ { "rfilename": "UD-IQ1_S/Qwen3.8-Flash-Next-UD-IQ1_S-00001-of-00003.gguf", "size": 10946624 }, { "rfilename": "UD-IQ1_S/Qwen3.8-Flash-Next-UD-IQ1_S-00002-of-00003.gguf", "size": 49990818368 }, { "rfilename": "UD-IQ1_S/Qwen3.8-Flash-Next-UD-IQ1_S-00003-of-00003.gguf", "size": 22544696352 } ], "hardware": "Unsloth MTP guide: about 76 GB total RAM+VRAM or unified memory for this bit tier [7]; context/OS headroom still matters. Strata SSD-paging can run with less, at a speed cost.", "kld": { "value": 0.39607, "corpus": "Unsloth guide; corpus/context not specified in displayed table", "context": null, "reference": "Qwen source", "source": 7, "top_token_agreement": "77.325%" }, "quality": "Top-token agreement: 77.325%; not benchmark task accuracy.", "decode": null, "memory_plan": { "total_gb": 76, "kind": "publisher approximate MTP memory by bit tier", "source": 7 } }, { "group": "normal-community", "family": "Normal Qwen / Unsloth GGUF", "tier": "UD-IQ1_M", "model_id": "unsloth/Qwen3.8-Flash-Next-GGUF", "source": 6, "size_bytes": 74538755776, "files": [ { "rfilename": "UD-IQ1_M/Qwen3.8-Flash-Next-UD-IQ1_M-00001-of-00003.gguf", "size": 10946624 }, { "rfilename": "UD-IQ1_M/Qwen3.8-Flash-Next-UD-IQ1_M-00002-of-00003.gguf", "size": 49988981792 }, { "rfilename": "UD-IQ1_M/Qwen3.8-Flash-Next-UD-IQ1_M-00003-of-00003.gguf", "size": 24538827360 } ], "hardware": "Unsloth MTP guide: about 76 GB total RAM+VRAM or unified memory for this bit tier [7]; context/OS headroom still matters. Strata SSD-paging can run with less, at a speed cost.", "kld": { "value": 0.314739, "corpus": "Unsloth guide; corpus/context not specified in displayed table", "context": null, "reference": "Qwen source", "source": 7, "top_token_agreement": "79.691%" }, "quality": "Top-token agreement: 79.691%; not benchmark task accuracy.", "decode": null, "memory_plan": { "total_gb": 76, "kind": "publisher approximate MTP memory by bit tier", "source": 7 } }, { "group": "normal-community", "family": "Normal Qwen / Unsloth GGUF", "tier": "UD-Q2_K_XL", "model_id": "unsloth/Qwen3.8-Flash-Next-GGUF", "source": 6, "size_bytes": 78869128864, "files": [ { "rfilename": "UD-Q2_K_XL/Qwen3.8-Flash-Next-UD-Q2_K_XL-00001-of-00003.gguf", "size": 10946624 }, { "rfilename": "UD-Q2_K_XL/Qwen3.8-Flash-Next-UD-Q2_K_XL-00002-of-00003.gguf", "size": 49979779296 }, { "rfilename": "UD-Q2_K_XL/Qwen3.8-Flash-Next-UD-Q2_K_XL-00003-of-00003.gguf", "size": 28878402944 } ], "hardware": "Unsloth MTP guide: about 80 GB total RAM+VRAM or unified memory for this bit tier [7]; context/OS headroom still matters. Strata SSD-paging can run with less, at a speed cost.", "kld": { "value": 0.224607, "corpus": "Unsloth guide; corpus/context not specified in displayed table", "context": null, "reference": "Qwen source", "source": 7, "top_token_agreement": "82.715%" }, "quality": "Top-token agreement: 82.715%; not benchmark task accuracy.", "decode": null, "memory_plan": { "total_gb": 80, "kind": "publisher approximate MTP memory by bit tier", "source": 7 } }, { "group": "normal-community", "family": "Normal Qwen / Unsloth GGUF", "tier": "UD-IQ3_XXS", "model_id": "unsloth/Qwen3.8-Flash-Next-GGUF", "source": 6, "size_bytes": 81961823936, "files": [ { "rfilename": "UD-IQ3_XXS/Qwen3.8-Flash-Next-UD-IQ3_XXS-00001-of-00003.gguf", "size": 10946624 }, { "rfilename": "UD-IQ3_XXS/Qwen3.8-Flash-Next-UD-IQ3_XXS-00002-of-00003.gguf", "size": 49567921344 }, { "rfilename": "UD-IQ3_XXS/Qwen3.8-Flash-Next-UD-IQ3_XXS-00003-of-00003.gguf", "size": 32382955968 } ], "hardware": "Unsloth MTP guide: about 91 GB total RAM+VRAM or unified memory for this bit tier [7]; context/OS headroom still matters. Strata SSD-paging can run with less, at a speed cost.", "kld": { "value": 0.16512, "corpus": "Unsloth guide; corpus/context not specified in displayed table", "context": null, "reference": "Qwen source", "source": 7, "top_token_agreement": "85.414%" }, "quality": "Top-token agreement: 85.414%; not benchmark task accuracy.", "decode": null, "memory_plan": { "total_gb": 91, "kind": "publisher approximate MTP memory by bit tier", "source": 7 } }, { "group": "normal-community", "family": "Normal Qwen / Unsloth GGUF", "tier": "UD-Q3_K_XL", "model_id": "unsloth/Qwen3.8-Flash-Next-GGUF", "source": 6, "size_bytes": 89986353824, "files": [ { "rfilename": "UD-Q3_K_XL/Qwen3.8-Flash-Next-UD-Q3_K_XL-00001-of-00003.gguf", "size": 10946624 }, { "rfilename": "UD-Q3_K_XL/Qwen3.8-Flash-Next-UD-Q3_K_XL-00002-of-00003.gguf", "size": 49983253824 }, { "rfilename": "UD-Q3_K_XL/Qwen3.8-Flash-Next-UD-Q3_K_XL-00003-of-00003.gguf", "size": 39992153376 } ], "hardware": "Unsloth MTP guide: about 91 GB total RAM+VRAM or unified memory for this bit tier [7]; context/OS headroom still matters. Strata SSD-paging can run with less, at a speed cost.", "kld": { "value": 0.106504, "corpus": "Unsloth guide; corpus/context not specified in displayed table", "context": null, "reference": "Qwen source", "source": 7, "top_token_agreement": "88.315%" }, "quality": "Top-token agreement: 88.315%; not benchmark task accuracy.", "decode": null, "memory_plan": { "total_gb": 91, "kind": "publisher approximate MTP memory by bit tier", "source": 7 } }, { "group": "normal-community", "family": "Normal Qwen / Unsloth GGUF", "tier": "UD-IQ4_XS", "model_id": "unsloth/Qwen3.8-Flash-Next-GGUF", "source": 6, "size_bytes": 93682584224, "files": [ { "rfilename": "UD-IQ4_XS/Qwen3.8-Flash-Next-UD-IQ4_XS-00001-of-00003.gguf", "size": 10946624 }, { "rfilename": "UD-IQ4_XS/Qwen3.8-Flash-Next-UD-IQ4_XS-00002-of-00003.gguf", "size": 49835229856 }, { "rfilename": "UD-IQ4_XS/Qwen3.8-Flash-Next-UD-IQ4_XS-00003-of-00003.gguf", "size": 43836407744 } ], "hardware": "Unsloth MTP guide: about 97 GB total RAM+VRAM or unified memory for this bit tier [7]; context/OS headroom still matters. Strata SSD-paging can run with less, at a speed cost.", "kld": { "value": 0.08363, "corpus": "Unsloth guide; corpus/context not specified in displayed table", "context": null, "reference": "Qwen source", "source": 7, "top_token_agreement": "89.554%" }, "quality": "Top-token agreement: 89.554%; not benchmark task accuracy.", "decode": null, "memory_plan": { "total_gb": 97, "kind": "publisher approximate MTP memory by bit tier", "source": 7 } }, { "group": "normal-community", "family": "Normal Qwen / Unsloth GGUF", "tier": "UD-Q4_K_XL", "model_id": "unsloth/Qwen3.8-Flash-Next-GGUF", "source": 6, "size_bytes": 111334654784, "files": [ { "rfilename": "UD-Q4_K_XL/Qwen3.8-Flash-Next-UD-Q4_K_XL-00001-of-00004.gguf", "size": 10946624 }, { "rfilename": "UD-Q4_K_XL/Qwen3.8-Flash-Next-UD-Q4_K_XL-00002-of-00004.gguf", "size": 49859583136 }, { "rfilename": "UD-Q4_K_XL/Qwen3.8-Flash-Next-UD-Q4_K_XL-00003-of-00004.gguf", "size": 49376141504 }, { "rfilename": "UD-Q4_K_XL/Qwen3.8-Flash-Next-UD-Q4_K_XL-00004-of-00004.gguf", "size": 12087983520 } ], "hardware": "Unsloth MTP guide: about 115 GB total RAM+VRAM or unified memory for this bit tier [7]; context/OS headroom still matters. Strata SSD-paging can run with less, at a speed cost.", "kld": { "value": 0.046893, "corpus": "Unsloth guide; corpus/context not specified in displayed table", "context": null, "reference": "Qwen source", "source": 7, "top_token_agreement": "92.255%" }, "quality": "Top-token agreement: 92.255%; not benchmark task accuracy.", "decode": { "tps_range": [ 7, 8.5 ], "source": "strata-q4", "hardware": "RTX 5070 12 GB; 64 GB RAM; SSD-paged experts, 40 GiB expert budget.", "notes": "Separate Strata workload/configuration; not comparable to short-chat GSQ table." }, "memory_plan": { "total_gb": 115, "kind": "publisher approximate MTP memory by bit tier", "source": 7 } }, { "group": "normal-community", "family": "Normal Qwen / Unsloth GGUF", "tier": "UD-Q5_K_XL", "model_id": "unsloth/Qwen3.8-Flash-Next-GGUF", "source": 6, "size_bytes": 158286406650, "files": [ { "rfilename": "UD-Q5_K_XL/Qwen3.8-Flash-Next-UD-Q5_K_XL-00001-of-00006.gguf", "size": 10946618 }, { "rfilename": "UD-Q5_K_XL/Qwen3.8-Flash-Next-UD-Q5_K_XL-00002-of-00006.gguf", "size": 682434912 }, { "rfilename": "UD-Q5_K_XL/Qwen3.8-Flash-Next-UD-Q5_K_XL-00003-of-00006.gguf", "size": 54400261312 }, { "rfilename": "UD-Q5_K_XL/Qwen3.8-Flash-Next-UD-Q5_K_XL-00004-of-00006.gguf", "size": 49990245824 }, { "rfilename": "UD-Q5_K_XL/Qwen3.8-Flash-Next-UD-Q5_K_XL-00005-of-00006.gguf", "size": 49882416192 }, { "rfilename": "UD-Q5_K_XL/Qwen3.8-Flash-Next-UD-Q5_K_XL-00006-of-00006.gguf", "size": 3320101792 } ], "hardware": "Unsloth MTP guide: about 164 GB total RAM+VRAM or unified memory for this bit tier [7]; context/OS headroom still matters. Strata SSD-paging can run with less, at a speed cost.", "kld": { "value": 0.030415, "corpus": "Unsloth guide; corpus/context not specified in displayed table", "context": null, "reference": "Qwen source", "source": 7, "top_token_agreement": "93.680%" }, "quality": "Top-token agreement: 93.680%; not benchmark task accuracy.", "decode": null, "memory_plan": { "total_gb": 164, "kind": "publisher approximate MTP memory by bit tier", "source": 7 } }, { "group": "normal-community", "family": "Normal Qwen / Unsloth GGUF", "tier": "UD-Q6_K_XL", "model_id": "unsloth/Qwen3.8-Flash-Next-GGUF", "source": 6, "size_bytes": 169165382688, "files": [ { "rfilename": "UD-Q6_K_XL/Qwen3.8-Flash-Next-UD-Q6_K_XL-00001-of-00006.gguf", "size": 10946624 }, { "rfilename": "UD-Q6_K_XL/Qwen3.8-Flash-Next-UD-Q6_K_XL-00002-of-00006.gguf", "size": 682434912 }, { "rfilename": "UD-Q6_K_XL/Qwen3.8-Flash-Next-UD-Q6_K_XL-00003-of-00006.gguf", "size": 54400261312 }, { "rfilename": "UD-Q6_K_XL/Qwen3.8-Flash-Next-UD-Q6_K_XL-00004-of-00006.gguf", "size": 49814108512 }, { "rfilename": "UD-Q6_K_XL/Qwen3.8-Flash-Next-UD-Q6_K_XL-00005-of-00006.gguf", "size": 49419317344 }, { "rfilename": "UD-Q6_K_XL/Qwen3.8-Flash-Next-UD-Q6_K_XL-00006-of-00006.gguf", "size": 14838313984 } ], "hardware": "Conservative full-file-resident plan (estimate): ~192 GB RAM+VRAM or unified memory. Weight bytes + 16 GB reserve rounded to 32 GB; not a tested minimum; paging/context alter needs.", "kld": { "value": 0.027091, "corpus": "Unsloth guide; corpus/context not specified in displayed table", "context": null, "reference": "Qwen source", "source": 7, "top_token_agreement": "94.089%" }, "quality": "Top-token agreement: 94.089%; not benchmark task accuracy.", "decode": null, "memory_plan": { "total_gb": 192, "kind": "derived conservative estimate, not minimum", "formula": "ceil((decimal weight GB + 16) / 32) * 32" } }, { "group": "normal-community", "family": "Normal Qwen / Unsloth GGUF", "tier": "Q8_0", "model_id": "unsloth/Qwen3.8-Flash-Next-GGUF", "source": 6, "size_bytes": 188225033248, "files": [ { "rfilename": "Q8_0/Qwen3.8-Flash-Next-Q8_0-00001-of-00006.gguf", "size": 10946624 }, { "rfilename": "Q8_0/Qwen3.8-Flash-Next-Q8_0-00002-of-00006.gguf", "size": 682434912 }, { "rfilename": "Q8_0/Qwen3.8-Flash-Next-Q8_0-00003-of-00006.gguf", "size": 54400261312 }, { "rfilename": "Q8_0/Qwen3.8-Flash-Next-Q8_0-00004-of-00006.gguf", "size": 49446841216 }, { "rfilename": "Q8_0/Qwen3.8-Flash-Next-Q8_0-00005-of-00006.gguf", "size": 49668930400 }, { "rfilename": "Q8_0/Qwen3.8-Flash-Next-Q8_0-00006-of-00006.gguf", "size": 34015618784 } ], "hardware": "Unsloth MTP guide: about 200 GB total RAM+VRAM or unified memory for this bit tier [7]; context/OS headroom still matters. Strata SSD-paging can run with less, at a speed cost.", "kld": { "value": 0.026574, "corpus": "Unsloth guide; corpus/context not specified in displayed table", "context": null, "reference": "Qwen source", "source": 7, "top_token_agreement": "94.122%" }, "quality": "Top-token agreement: 94.122%; not benchmark task accuracy.", "decode": null, "memory_plan": { "total_gb": 200, "kind": "publisher approximate MTP memory by bit tier", "source": 7 } }, { "group": "accelerators", "family": "Swift 1.5 non-GGUF", "tier": "AWQ W4A16", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-W4A16-AWQ", "source": 10, "size_bytes": 177249648728, "files": [ { "rfilename": "embedding-model-00006-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00007-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00008-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00009-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00010-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00011-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00012-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00013-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00014-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00015-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00016-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00017-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00018-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00019-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00020-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00021-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00022-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00023-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00024-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00025-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00026-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00027-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00028-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00029-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00030-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00031-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "embedding-model-00032-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "embedding-model-00033-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "embedding-model-00034-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "embedding-model-00035-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "embedding-model-00036-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "model-00001-of-00019.safetensors", "size": 4001006440 }, { "rfilename": "model-00002-of-00019.safetensors", "size": 3200016032 }, { "rfilename": "model-00003-of-00019.safetensors", "size": 4002440096 }, { "rfilename": "model-00004-of-00019.safetensors", "size": 4002411080 }, { "rfilename": "model-00005-of-00019.safetensors", "size": 3997150720 }, { "rfilename": "model-00006-of-00019.safetensors", "size": 4001768592 }, { "rfilename": "model-00007-of-00019.safetensors", "size": 4002396256 }, { "rfilename": "model-00008-of-00019.safetensors", "size": 4002073608 }, { "rfilename": "model-00009-of-00019.safetensors", "size": 4002396112 }, { "rfilename": "model-00010-of-00019.safetensors", "size": 4002396240 }, { "rfilename": "model-00011-of-00019.safetensors", "size": 4002396240 }, { "rfilename": "model-00012-of-00019.safetensors", "size": 4002073600 }, { "rfilename": "model-00013-of-00019.safetensors", "size": 4002396112 }, { "rfilename": "model-00014-of-00019.safetensors", "size": 4002396608 }, { "rfilename": "model-00015-of-00019.safetensors", "size": 4002396712 }, { "rfilename": "model-00016-of-00019.safetensors", "size": 4002503144 }, { "rfilename": "model-00017-of-00019.safetensors", "size": 4002396376 }, { "rfilename": "model-00018-of-00019.safetensors", "size": 4002396240 }, { "rfilename": "model-00019-of-00019.safetensors", "size": 1601836056 }, { "rfilename": "model_mtp.safetensors", "size": 5214305424 } ], "hardware": "Compatible vLLM/runtime and capacity for listed weight files plus cache; no tested minimum found.", "kld": null, "quality": "No quant-specific task score or KLD located; card repeats BF16 benchmarks.", "decode": null }, { "group": "accelerators", "family": "Swift 1.5 non-GGUF", "tier": "AutoRound W4A16", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-W4A16-AutoRound", "source": 11, "size_bytes": 177237111728, "files": [ { "rfilename": "model-00001-of-00081.safetensors", "size": 1377483080 }, { "rfilename": "model-00002-of-00081.safetensors", "size": 2951071312 }, { "rfilename": "model-00003-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00004-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00005-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00006-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00007-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00008-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00009-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00010-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00011-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00012-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00013-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00014-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00015-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00016-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00017-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00018-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00019-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00020-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00021-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00022-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00023-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00024-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00025-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00026-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00027-of-00081.safetensors", "size": 3200016032 }, { "rfilename": "model-00028-of-00081.safetensors", "size": 3200016040 }, { "rfilename": "model-00029-of-00081.safetensors", "size": 3200016040 }, { "rfilename": "model-00030-of-00081.safetensors", "size": 3200016040 }, { "rfilename": "model-00031-of-00081.safetensors", "size": 3200016040 }, { "rfilename": "model-00032-of-00081.safetensors", "size": 3200016040 }, { "rfilename": "model-00033-of-00081.safetensors", "size": 3200016040 }, { "rfilename": "model-00034-of-00081.safetensors", "size": 1692108480 }, { "rfilename": "model-00035-of-00081.safetensors", "size": 1377483080 }, { "rfilename": "model-00036-of-00081.safetensors", "size": 1376101696 }, { "rfilename": "model-00037-of-00081.safetensors", "size": 1377483080 }, { "rfilename": "model-00038-of-00081.safetensors", "size": 1377483080 }, { "rfilename": "model-00039-of-00081.safetensors", "size": 1377483080 }, { "rfilename": "model-00040-of-00081.safetensors", "size": 1376101696 }, { "rfilename": "model-00041-of-00081.safetensors", "size": 1377483080 }, { "rfilename": "model-00042-of-00081.safetensors", "size": 1377483080 }, { "rfilename": "model-00043-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00044-of-00081.safetensors", "size": 1376106336 }, { "rfilename": "model-00045-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00046-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00047-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00048-of-00081.safetensors", "size": 1376106336 }, { "rfilename": "model-00049-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00050-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00051-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00052-of-00081.safetensors", "size": 1376106336 }, { "rfilename": "model-00053-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00054-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00055-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00056-of-00081.safetensors", "size": 1376106336 }, { "rfilename": "model-00057-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00058-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00059-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00060-of-00081.safetensors", "size": 1376106336 }, { "rfilename": "model-00061-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00062-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00063-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00064-of-00081.safetensors", "size": 1376106336 }, { "rfilename": "model-00065-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00066-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00067-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00068-of-00081.safetensors", "size": 1376106336 }, { "rfilename": "model-00069-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00070-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00071-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00072-of-00081.safetensors", "size": 1376106336 }, { "rfilename": "model-00073-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00074-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00075-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00076-of-00081.safetensors", "size": 1376106336 }, { "rfilename": "model-00077-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00078-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00079-of-00081.safetensors", "size": 1377487720 }, { "rfilename": "model-00080-of-00081.safetensors", "size": 1376106336 }, { "rfilename": "model-00081-of-00081.safetensors", "size": 3453825000 }, { "rfilename": "model_mtp.safetensors", "size": 5214305424 } ], "hardware": "Compatible vLLM/runtime and capacity for listed weight files plus cache; no tested minimum found.", "kld": null, "quality": "No quant-specific task score or KLD located; card repeats BF16 benchmarks.", "decode": null }, { "group": "accelerators", "family": "Swift 1.5 non-GGUF", "tier": "NVFP4", "model_id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-NVFP4", "source": 12, "size_bytes": 186395943704, "files": [ { "rfilename": "embedding-mixed.safetensors", "size": 3200016000 }, { "rfilename": "embedding-model-00006-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00007-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00008-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00009-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00010-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00011-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00012-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00013-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00014-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00015-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00016-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00017-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00018-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00019-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00020-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00021-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00022-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00023-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00024-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00025-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00026-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00027-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00028-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00029-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00030-of-00131.safetensors", "size": 3200016032 }, { "rfilename": "embedding-model-00031-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "embedding-model-00032-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "embedding-model-00033-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "embedding-model-00034-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "embedding-model-00035-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "embedding-model-00036-of-00131.safetensors", "size": 3200016040 }, { "rfilename": "model-layer-00000.safetensors", "size": 1571201568 }, { "rfilename": "model-layer-00001.safetensors", "size": 1636882816 }, { "rfilename": "model-layer-00002.safetensors", "size": 1571201568 }, { "rfilename": "model-layer-00003.safetensors", "size": 1558177344 }, { "rfilename": "model-layer-00004.safetensors", "size": 1571201568 }, { "rfilename": "model-layer-00005.safetensors", "size": 1571201568 }, { "rfilename": "model-layer-00006.safetensors", "size": 1571201568 }, { "rfilename": "model-layer-00007.safetensors", "size": 1558177344 }, { "rfilename": "model-layer-00008.safetensors", "size": 1571201568 }, { "rfilename": "model-layer-00009.safetensors", "size": 1571201568 }, { "rfilename": "model-layer-00010.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00011.safetensors", "size": 1558183512 }, { "rfilename": "model-layer-00012.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00013.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00014.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00015.safetensors", "size": 1558183512 }, { "rfilename": "model-layer-00016.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00017.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00018.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00019.safetensors", "size": 1558183512 }, { "rfilename": "model-layer-00020.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00021.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00022.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00023.safetensors", "size": 1558183512 }, { "rfilename": "model-layer-00024.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00025.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00026.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00027.safetensors", "size": 1558183512 }, { "rfilename": "model-layer-00028.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00029.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00030.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00031.safetensors", "size": 1558183512 }, { "rfilename": "model-layer-00032.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00033.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00034.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00035.safetensors", "size": 1558183512 }, { "rfilename": "model-layer-00036.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00037.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00038.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00039.safetensors", "size": 1558183512 }, { "rfilename": "model-layer-00040.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00041.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00042.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00043.safetensors", "size": 1558183512 }, { "rfilename": "model-layer-00044.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00045.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00046.safetensors", "size": 1571207736 }, { "rfilename": "model-layer-00047.safetensors", "size": 1558183512 }, { "rfilename": "model-tail.safetensors", "size": 8668130456 } ], "hardware": "Native NVFP4 needs compatible NVIDIA Blackwell runtime; listed weight size exceeds one 24 GB GPU. Add cache/runtime headroom.", "kld": null, "quality": "No quant-specific task score or KLD located; card repeats BF16 benchmarks.", "decode": null } ], "bf16_benchmarks": [ { "benchmark": "GPQA-Diamond", "base": "89.80%", "swift": "89.60%", "base_tokens": "17,683", "swift_tokens": "7,823", "source": 2 }, { "benchmark": "MMLU-Pro", "base": "87.75%", "swift": "87.20%", "base_tokens": "3,528", "swift_tokens": "1,519", "source": 2 }, { "benchmark": "IFBench", "base": "73.20%", "swift": "70.13%", "base_tokens": "8,310", "swift_tokens": "4,411", "source": 2 }, { "benchmark": "LiveCodeBench v6", "base": "88.40%", "swift": "90.39%", "base_tokens": "17,833", "swift_tokens": "9,849", "source": 2 }, { "benchmark": "Terminal-Bench 2.1*", "base": "67.64%", "swift": "69.66%", "base_tokens": "40,591", "swift_tokens": "45,428", "source": 2 } ], "gsq_kld_reporting": [ { "tier": "IQ2_XS", "domain": "prose_c4", "kld": "0.188117", "reported_error": "0.001955", "chunks": "100", "context": "512", "ista_kld": "0.200267", "starter_kld": "0.215863", "delta_vs_ista_pct": "-6.066900687582077", "delta_vs_starter_pct": "-12.853522836243359" }, { "tier": "IQ2_XS", "domain": "code_codeparrot", "kld": "0.174294", "reported_error": "0.002979", "chunks": "100", "context": "512", "ista_kld": "0.184403", "starter_kld": "0.193624", "delta_vs_ista_pct": "-5.4820149346811125", "delta_vs_starter_pct": "-9.983266537206125" }, { "tier": "IQ2_XS", "domain": "math_gsm8k", "kld": "0.120991", "reported_error": "0.002643", "chunks": "100", "context": "512", "ista_kld": "0.132848", "starter_kld": "0.146175", "delta_vs_ista_pct": "-8.925237865831626", "delta_vs_starter_pct": "-17.228664272276383" }, { "tier": "IQ2_XS", "domain": "multilingual_mc4_de", "kld": "0.166814", "reported_error": "0.00467", "chunks": "25", "context": "512", "ista_kld": "0.158851", "starter_kld": "0.185316", "delta_vs_ista_pct": "5.012873699252762", "delta_vs_starter_pct": "-9.984027283127206" }, { "tier": "IQ2_XS", "domain": "multilingual_mc4_fr", "kld": "0.213379", "reported_error": "0.005505", "chunks": "25", "context": "512", "ista_kld": "0.227306", "starter_kld": "0.241407", "delta_vs_ista_pct": "-6.126983009687381", "delta_vs_starter_pct": "-11.610268136383784" }, { "tier": "IQ2_XS", "domain": "multilingual_mc4_es", "kld": "0.119114", "reported_error": "0.004094", "chunks": "25", "context": "512", "ista_kld": "0.126496", "starter_kld": "0.135836", "delta_vs_ista_pct": "-5.835757652415885", "delta_vs_starter_pct": "-12.310433169410185" }, { "tier": "IQ2_XS", "domain": "multilingual_mc4_zh", "kld": "0.264923", "reported_error": "0.005239", "chunks": "25", "context": "512", "ista_kld": "0.296037", "starter_kld": "0.289135", "delta_vs_ista_pct": "-10.510172714897115", "delta_vs_starter_pct": "-8.373942967817783" }, { "tier": "IQ2_XS", "domain": "fresh_c4_en_shard6", "kld": "0.186271", "reported_error": "0.001944", "chunks": "100", "context": "512", "ista_kld": "0.198573", "starter_kld": "", "delta_vs_ista_pct": "-6.195202771776631", "delta_vs_starter_pct": "" }, { "tier": "Q2_0", "domain": "code_codeparrot", "kld": "0.235256", "reported_error": "0.003868", "chunks": "100", "context": "512", "ista_kld": "0.219545", "starter_kld": "0.228021", "delta_vs_ista_pct": "7.156163884397282", "delta_vs_starter_pct": "3.1729533683301003" }, { "tier": "Q2_0", "domain": "fresh_c4_en_shard6", "kld": "0.236228", "reported_error": "0.002652", "chunks": "100", "context": "512", "ista_kld": "0.218501", "starter_kld": "", "delta_vs_ista_pct": "8.113006347797036", "delta_vs_starter_pct": "" }, { "tier": "Q2_0", "domain": "math_gsm8k", "kld": "0.149809", "reported_error": "0.003194", "chunks": "100", "context": "512", "ista_kld": "0.163066", "starter_kld": "0.202867", "delta_vs_ista_pct": "-8.129836998515938", "delta_vs_starter_pct": "-26.15408124534794" }, { "tier": "Q2_0", "domain": "multilingual_mc4_de", "kld": "0.219444", "reported_error": "0.00608", "chunks": "25", "context": "512", "ista_kld": "0.192854", "starter_kld": "0.248268", "delta_vs_ista_pct": "13.78763209474525", "delta_vs_starter_pct": "-11.610034317753392" }, { "tier": "Q2_0", "domain": "multilingual_mc4_es", "kld": "0.148757", "reported_error": "0.004586", "chunks": "25", "context": "512", "ista_kld": "0.144779", "starter_kld": "0.171527", "delta_vs_ista_pct": "2.7476360521899057", "delta_vs_starter_pct": "-13.274878007544011" }, { "tier": "Q2_0", "domain": "multilingual_mc4_fr", "kld": "0.300467", "reported_error": "0.007617", "chunks": "25", "context": "512", "ista_kld": "0.2717", "starter_kld": "0.337115", "delta_vs_ista_pct": "10.587780640412214", "delta_vs_starter_pct": "-10.871067736529083" }, { "tier": "Q2_0", "domain": "multilingual_mc4_zh", "kld": "0.385924", "reported_error": "0.007076", "chunks": "25", "context": "512", "ista_kld": "0.446938", "starter_kld": "0.376247", "delta_vs_ista_pct": "-13.651557934210112", "delta_vs_starter_pct": "2.5719806403772028" }, { "tier": "Q2_0", "domain": "prose_c4", "kld": "0.234242", "reported_error": "0.00257", "chunks": "100", "context": "512", "ista_kld": "0.218335", "starter_kld": "0.270309", "delta_vs_ista_pct": "7.285593239746269", "delta_vs_starter_pct": "-13.34287796558754" }, { "tier": "IQ3_XXS", "domain": "prose_c4", "kld": "0.116077", "reported_error": "0.001601", "chunks": "100", "context": "512", "ista_kld": "0.114982", "starter_kld": "0.125842", "delta_vs_ista_pct": "0.9523229722913085", "delta_vs_starter_pct": "-7.7597304556507485" }, { "tier": "IQ3_XXS", "domain": "code_codeparrot", "kld": "0.118707", "reported_error": "0.002266", "chunks": "100", "context": "512", "ista_kld": "0.117015", "starter_kld": "0.122811", "delta_vs_ista_pct": "1.4459684655813376", "delta_vs_starter_pct": "-3.3417202042162475" }, { "tier": "IQ3_XXS", "domain": "math_gsm8k", "kld": "0.086821", "reported_error": "0.002149", "chunks": "100", "context": "512", "ista_kld": "0.094208", "starter_kld": "0.103453", "delta_vs_ista_pct": "-7.841159986413048", "delta_vs_starter_pct": "-16.076865823127417" }, { "tier": "IQ3_XXS", "domain": "multilingual_mc4_de", "kld": "0.1091", "reported_error": "0.003632", "chunks": "25", "context": "512", "ista_kld": "0.104781", "starter_kld": "0.12331", "delta_vs_ista_pct": "4.12193050266747", "delta_vs_starter_pct": "-11.523801800340605" }, { "tier": "IQ3_XXS", "domain": "multilingual_mc4_fr", "kld": "0.133624", "reported_error": "0.003993", "chunks": "25", "context": "512", "ista_kld": "0.130922", "starter_kld": "0.152166", "delta_vs_ista_pct": "2.063824261774183", "delta_vs_starter_pct": "-12.18537649672069" }, { "tier": "IQ3_XXS", "domain": "multilingual_mc4_es", "kld": "0.073681", "reported_error": "0.002812", "chunks": "25", "context": "512", "ista_kld": "0.073858", "starter_kld": "0.080745", "delta_vs_ista_pct": "-0.23964905629721978", "delta_vs_starter_pct": "-8.748529320700971" }, { "tier": "IQ3_XXS", "domain": "multilingual_mc4_zh", "kld": "0.174102", "reported_error": "0.004177", "chunks": "25", "context": "512", "ista_kld": "0.183406", "starter_kld": "0.169308", "delta_vs_ista_pct": "-5.072898378460899", "delta_vs_starter_pct": "2.831525976327187" }, { "tier": "IQ3_S", "domain": "prose_c4", "kld": "0.062556", "reported_error": "0.000816", "chunks": "100", "context": "512", "ista_kld": "0.062341", "starter_kld": "0.07367", "delta_vs_ista_pct": "0.34487736802425356", "delta_vs_starter_pct": "-15.086195194787566" }, { "tier": "IQ3_S", "domain": "code_codeparrot", "kld": "0.069016", "reported_error": "0.001395", "chunks": "100", "context": "512", "ista_kld": "0.069075", "starter_kld": "0.078276", "delta_vs_ista_pct": "-0.08541440463265074", "delta_vs_starter_pct": "-11.82993510143595" }, { "tier": "IQ3_S", "domain": "math_gsm8k", "kld": "0.05496", "reported_error": "0.001441", "chunks": "100", "context": "512", "ista_kld": "0.057079", "starter_kld": "0.065384", "delta_vs_ista_pct": "-3.712398605441574", "delta_vs_starter_pct": "-15.942738284595613" }, { "tier": "IQ3_S", "domain": "multilingual_mc4_de", "kld": "0.063634", "reported_error": "0.002354", "chunks": "25", "context": "512", "ista_kld": "0.061909", "starter_kld": "0.075633", "delta_vs_ista_pct": "2.7863477038879605", "delta_vs_starter_pct": "-15.864768024539563" }, { "tier": "IQ3_S", "domain": "multilingual_mc4_fr", "kld": "0.07676", "reported_error": "0.002782", "chunks": "25", "context": "512", "ista_kld": "0.070573", "starter_kld": "0.092368", "delta_vs_ista_pct": "8.766808836240486", "delta_vs_starter_pct": "-16.89762688376928" }, { "tier": "IQ3_S", "domain": "multilingual_mc4_es", "kld": "0.042848", "reported_error": "0.002029", "chunks": "25", "context": "512", "ista_kld": "0.043018", "starter_kld": "0.048054", "delta_vs_ista_pct": "-0.3951834115951543", "delta_vs_starter_pct": "-10.8336454821659" }, { "tier": "IQ3_S", "domain": "multilingual_mc4_zh", "kld": "0.095915", "reported_error": "0.002657", "chunks": "25", "context": "512", "ista_kld": "0.103355", "starter_kld": "0.10152", "delta_vs_ista_pct": "-7.198490639059554", "delta_vs_starter_pct": "-5.521079590228525" }, { "tier": "IQ3_S", "domain": "fresh_c4_en_shard6", "kld": "0.067801", "reported_error": "0.001056", "chunks": "100", "context": "512", "ista_kld": "0.067419", "starter_kld": "0.076638", "delta_vs_ista_pct": "0.5666058529494555", "delta_vs_starter_pct": "-11.530833268091543" } ] }