Mistral Research Head on Le Chonk: Trained on 3,800 GPUs in Mistral's Own Datacenters
mattturck · x · 2026-10-26
Matt Turck interviews Pierre Stock, who leads research at Mistral, on the company's comeback and "Le Chonk," billed as the West's biggest open-weight model.
Key points:
- The Western open-weight comeback: Beam, Le Chonk, Muse, Nemotron
- Le Chonk was trained on 3,800 GPUs in Mistral's own datacenters
- Meta keeps its best model closed while Mistral opens Chonk; comparison vs Beam, Muse Spark Open and Chinese models
- Why Chinese labs remain at the frontier; the US shutdown of Anthropic's models ("AI is like electricity")
- Stock's path: FAIR, Yann LeCun, and the "Mistral mafia"
More from Companies & People
- AI helped a game studio go from 4 games a year to 4 games a month — majidmanzarpour · 2026-10-26
- Musk talked AI for 40 minutes on Isaacson's podcast without mentioning LLMs — FinanceYF5 · 2026-10-26
- Slack founder memoir reveals Salesforce waste: Hawaii exec event cost millions — SumitGup · 2026-10-26
- Tesla's Texas robotaxi fleet nears 1,000 vehicles, up 78% in October — mitchdeg · 2026-10-26
- Security stays reactive in AI: CISOs can't get budgets past the board — annetgriffin · 2026-10-26
- Anthropic blocks distillation: it was 5-7% of revenue, tens of thousands of accounts banned — FinanceYF5 · 2026-10-26