Post-Cutoff

Arthur Mensch: “Trained and served on our own compute, and RL shows no sign of saturation”

Arthur Mensch @arthurmenschX

Why it matters

First-hand comment from Mistral’s CEO on the Mistral Large 4 launch (~111k views by 7 Oct).

Summary

Posted 13:29 UTC on 6 Oct, quoting Mistral’s launch post: “That one took some groundwork. Trained and served on our own compute, and RL shows no sign of saturation”. This matches Guillaume Lample’s thread, which says the preview comes from an RL run that is still in progress (2026-10-06-guillaumelample-ml4-thread). Reach at 11:31 UTC on 7 Oct: ~111k views, 3,350 likes.

Archived text

That one took some groundwork. Trained and served on our own compute, and RL shows no sign of saturation

Quoting @MistralAI: Meet Mistral Large 4, aka Le Chonk.

• 1T parameters, natively multimodal. 49B active. It is the best open weights model from US or Europe on aggregated benchmarks. • State-of-the-art on critical workloads, including cyber defense, manufacturing and finance and it surpasses closed frontier models on visual grounding. • Forged in Europe end-to-end and is deployable from Europe via our own Mistral Cloud infrastructure. • Available to all via API today. Working with cybersecurity partners privately.

Open weights release end of October.

Source: x.com/arthurmensch/status/2107463276326162840Archived 2026-10-07 via manual (api.fxtwitter.com, fetched 11:31 UTC).Counts: 110,849 views, 3,350 likes, 221 reposts, 132 replies, 25 quotes (at fetch time)

Cited in

  1. Model releases 98 days after the cutoff

    Mistral releases Mistral Large 4 (“le Chonk”)

People in this post

Guillaume Lample, Co-founder and Chief Scientist, Mistral AI