FreeLLMMistral releases open‑weight 1‑trillion‑parameter model Le Chonk with coding and security gains
Mistral Large 4 (Le Chonk) is now publicly available with three reasoning modes and top open‑weight coding performance.
Mistral announced that the weights for its 1‑trillion‑parameter multimodal model, Mistral Large 4 (Le Chonk), are now publicly released.
The model shows leading results on open‑weight coding benchmarks, scoring 19.7% on FrontierCode 2.0 and 41% on Terminal‑Bench 4.0, while using fewer tokens than competitors.
Three reasoning modes are offered: Low for simple tasks, Medium for everyday use, and High (default) for demanding reasoning, letting users balance performance and cost.
In cybersecurity tests, Le Chonk achieved a 51% score on the Artificial Analysis Cyber Index, indicating strong vulnerability discovery and patching abilities without excessive refusals.
The release follows a public preview that already placed the model at the top of open‑weight coding rankings, and it joins a landscape where the top free chat models are Kimi K3, GLM‑5.3, and MiMo‑V2.6‑Pro.
あなたにとっての意味
Free‑model users gain a new benchmark‑leading option that can be run locally, expanding the pool of capable, cost‑effective AI models.
FreeLLMArena上
Arenaのブラインド投票に基づく、31年10月2026日のランキングから：
- Mistral Large (2411): #116 among free chat models
ソース
- Le Chonk: Built for Coders, Trusted by Defenders., Mistral (official), 29 Oct 2026
上記のソースから自動的に生成し、当社のコードでチェックしました：すべての数値と名前が含まれ、文がソースをコピーしてはいけません。間違いを見つけましたか？報告する。
今日無料で使えるAIはどれですか？
すべての無料AIモデルをブラインドヒューマン投票で順位付けし、あなたのGPUで必要なメモリと速度を示します。