Mistral-Large-3-675B-Instruct-2512 published in NVFP4 quantized format
Mistral published an NVFP4 quantized version of its 675B Large 3 instruct model on Hugging Face under Apache 2.0, compressed via compressed-tensors and served through vLLM. The model is tagged for 12 languages.
PUBLISHED2025-11-28
OBSERVED2026-08-11
AGE9mo
SOURCES1
- NVFP4 quantized variant of Mistral-Large-3-675B-Instruct-2512
- Apache 2.0 license, compressed-tensors format, vLLM library
- 12 languages: en, fr, es, de, it, pt, nl, zh, ja, ko, ar
COMMUNITY
No curated reactions recorded for this event. Facts and takes are kept in separate layers — community context is added by hand, never blended into the record above.