- Mistral AI launched Large 4 on Oct. 6, a 1-trillion-parameter model that activates 49 billion parameters per query at $1.36 per million input tokens and $4.18 per million output tokens.
- The open-weight model, nicknamed “Le Chonk” after the “Le Chaton Fat” meme, will have its weights released by the end of October.
- On Finance Agent v2, Large 4 scores 54.7, ahead of GPT-6 Astra at 53.5 but behind Claude Opus 5.5 at 58.6.
- Mistral raised €3 billion ($3.37 billion) in a September Series D led by Samsung at a valuation above $23.6 billion.
Paris-based Mistral AI launched Mistral Large 4 on Tuesday, a 1-trillion-parameter model that uses a mixture-of-experts design to activate just 49 billion of its parameters per query. Chief scientist Guillaume Lample called it “at the frontier of open models, and by far the strongest open-weight model from the US or Europe“ on X.
The nickname “Le Chonk” traces back to a June meme. After Mistral renamed Le Chat to Vibe, fans invented a fictional model called Le Chaton Fat, prompting CEO Arthur Mensch to play along on X, and the launch post now lists the model as “very officially” le Chonk.
Mistral sells “sovereign AI,” letting countries and companies run models without handing data to outside firms. Saudi Arabia‘s HUMAIN signed a deal worth hundreds of millions of euros in August, and Mistral raised €3 billion ($3.37 billion) in a Samsung-led Series D in September at a valuation above €21 billion ($23.6 billion).
Meanwhile, Large 4 costs $1.36 per million input tokens and $4.18 per million output tokens — roughly a third and a fifth of Claude Opus 5.5’s prices, and about a seventh and a twelfth of GPT-6 Astra’s.
However, benchmarks tell a mixed story. On DeepSWE 1.1, Large 4 scored 62, topping GLM-5.3 and DeepSeek V4 Pro but trailing Kimi K3’s 68, per Datacurve’s leaderboard.
On AutomationBench, it scored 59.9, behind Claude Sonnet 5.5’s 71.8 and Opus 5.5’s 69.5. In a blind human coding evaluation by Surge AI, Large 4 ranked second with 3.74 out of 5, behind Claude Opus 5’s 4.22.
Mistral says it will release the model’s weights by the end of October, letting outside developers test those claims.
✅ Follow BITNEWSBOT on Telegram, Facebook, LinkedIn, X.com, and Google News for instant updates.
