French AI company Mistral has released the final version of its biggest model, nicknamed Le Chonk, for anyone to download, CNET and Wired report. Its formal name is Mistral Large 4, CNET reports. An early version came out in public preview earlier this month, the outlets report.

Mistral now says Le Chonk is the "world's best" open model for automated coding and cyber defense, Wired reports. The company says it fed the model harder and harder coding and security problems, the outlet reports. That let it match or beat models from top Chinese labs like DeepSeek, Mistral says.

"The response to the preview has exceeded our expectations," Mistral said in its post, CNET reports. The company says the model can read text as well as images, the outlet reports. It is built to check logs, review code and look for new security holes.

"It's a pivotal moment in the global AI leadership race," Mistral science chief Pierre Stock said at a press briefing, Wired reports. Stock argued that the West needs its own open models. "You have no guarantee that [Chinese models] won't behave maliciously," he said.

Mistral pitches open models as a tool that cannot be switched off from afar, unlike closed models from firms like OpenAI and Google, Wired reports. "AI is going to be like electricity," Stock said. "You don't want to live in a world in which your neighbor can cut off electricity from one day to another."

What we know: Mistral has released Le Chonk to the public and says it leads open models in coding and security.
What we don't: How it does in tests run by people outside the company.