AINewsnow

NVFP4 Lands: 119B Open Weights Now Fit in 66 GiB

This story is from 2026-09-24. It is preserved in the archive; the latest stories are on the live feed.

Originally published on AI Tech Connect . What actually changed This is not a model release. Mistral Small 4 shipped on 16 March 2026. The weights are six months old. Only the NVFP4 checkpoint activity is recent. The official NVFP4 checkpoint exists. mistralai/Mistral-Small-4-119B-2603-NVFP4 is pub…

Read the full story at DEV Community — Machine Learning ↗

Timeline · 1 report

  1. 2026-09-24 07:32 · DEV Community — Machine Learning
    NVFP4 Lands: 119B Open Weights Now Fit in 66 GiB

More stories

  1. Mistral Small: 58 empty replies in 300 tool calls, and the retry loop — DEV Community — AI
  2. Zero-Downtime LLM Inference on Kubernetes Spot Instances (~70% GPU Cost Savings) — DEV Community — Machine Learning
  3. AI Weekly — 2026-10-30 to 2026-11-06 | Open Weights Meet Production Constraints — DEV Community — AI
  4. Mistral adds managed agents and apps to its Studio platform — DEV Community — AI
  5. Mistral unveils open-weight coding agent model at top coding agent conference — Latent Space
  6. Config-Verified Comparison Maps How 10 Open MoE Models Route Tokens — MarkTechPost
  7. 2026 Guide Explains MoE Routing Strategies From Top-K to Dynamic Routing — MarkTechPost
  8. MIT study shows AI fixing AI, alongside Mistral open model and Google Genie 4 — Mixture of Experts (IBM)

Get the daily brief of stories like this at 6:30 every morning →