Weekly ranking, Technology

Mistral Large 4 (Le Chonk) is now open weight | Mistral

mistral.ai, Technology

Mistral's own release notes give concrete benchmark, safety, license and pricing details, but nearly all performance claims are self-reported and unverified.

What it adds

A vendor launch announcement listing benchmark results, safety measures, licensing and pricing for the open-weight release. Its distinctive elements are the token-efficiency framing and the admission of live-traffic cyber uplift.

Claim by claim

• 19 the author's own
• 1 an AI would say
• 1 known elsewhere
• 0 common knowledge

Its strongest original claims

• Mistral Large 4 ranks as the best open-weight model for coding.
• Mistral Large 4 leads Cognition's FrontierCode 2.0 benchmark with 19.7%.
• It scores 48% on the Artificial Analysis Coding Agent Index, tied with GLM-5.3 while using 47% fewer tokens.
• On Terminal-Bench 4.0 it improved 13 points to 41% with 30% fewer output tokens.
• Three weeks of additional reinforcement learning drove the coding gains, with no signs of saturation.

Scored by Expulp Verdict for what the piece adds beyond what is already known and published. The full verdict shows every claim with its reasons. Leaderboard, what Verdict gets right and where it is weak.