Changes / Feeds / Qwen/Qwen3.8-27B-FP8
Qwen/Qwen3.8-27B-FP8
3 changes to a tracked key since 2026-08-13, newest first. Head as last read: 017b9c7af6b5689d5dd426a76e0bc077eb5ca20a.
A quantization of Qwen/Qwen3.8-27B: 407 layers re-encoded; of the 792 tensors left unquantized, 792 byte-identical to the base. Compared tensor by tensor from sha256 fingerprints: 017b9c7af6b5 against Qwen/Qwen3.8-27B at 1d4bf0f2ff60. A hash is equal or not; it says nothing about what an edit does.
| Date | Advisory | What changed, and what it does |
|---|---|---|
| 2026-08-13 | x_NXMC-2026-0132 | generation defaults changed (temperature 0.6 to 1); a call that does not pass these explicitly now generates differently (2026-08-13).e4f7558fd850 -> 5c9be11f91f4 Open in app |
| 2026-08-13 | x_NXMC-2026-0131 breaking | the chat template changed and renders differently (system+user, multi-turn, tool-call); prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-08-13).806d12210bcd -> e4f7558fd850 Open in app |
| 2026-08-13 | x_NXMC-2026-0130 | tokenizer model_max_length 131072 to 262144; truncation defaults change (2026-08-13).10b09acb2fdc -> 806d12210bcd Open in app |
Pin it, and let a bot move the pin
from_pretrained("Qwen/Qwen3.8-27B-FP8", revision="017b9c7af6b5689d5dd426a76e0bc077eb5ca20a")With the Renovate preset, a pull request moves this pin when the head moves, and links back here.
[](https://www.neurarch.com/changes-feeds/models/Qwen/Qwen3.8-27B-FP8.html)