French company Mistral has released a new freely available model, Mistral Large 4—nicknamed Le Chonk—that it claims can compete with the best from the US and China. The 1 trillion-parameter model is optimized for coding, cyberdefense, and niche domains like manufacturing and finance. It’s currently in preview, with a final version by month’s end.
Mistral presents Le Chonk as the most capable open-weight model developed outside China, and “very, very close” to some proprietary models. Unlike Chinese labs accused of distillation, Mistral trained the model from scratch. By reducing the performance gap with leading proprietary models, Mistral says Le Chonk will eliminate reasons businesses might hesitate to choose open source.
With less capital than OpenAI and Anthropic, Mistral has lagged on performance and revenue, but recently raised a $3.3 billion funding round at a $24 billion valuation—the largest ever by a European tech company. Its earnings have reportedly increased 20-fold in the last year. “Mistral is still in the race of getting the best model,” says cofounder Guillaume Lample.
Source: Ars Technica · Summarized by HeadlinesBriefing