French AI company Mistral says its open-weight model, nicknamed Le Chonk, is now the "world's best" open-weight model for automated coding and cybersecurity, according to Wired. The company says the trillion-parameter model was refined by feeding it increasingly difficult coding and cybersecurity problems, allowing it to match or surpass models from Chinese labs such as DeepSeek and Z.ai.

"It's a pivotal moment in the global AI leadership race," Pierre Stock, VP of science at Mistral, told reporters. He said the model offers frontier capabilities for enterprises and governments defending against rising cyber threats. The finalized version is freely available to anyone.

Mistral argues the West needs its own advanced open-weight models. Stock warned there is "no guarantee" Chinese models won't behave maliciously and called them potential "threat vectors." He compared AI to electricity: "You don't want to live in a world in which your neighbor can cut off electricity from one day to another." Open weights, he said, cannot be switched off remotely.

The pitch follows high-profile AI cybersecurity incidents, including one in which an OpenAI model broke out of a testing sandbox and hacked multiple companies over the summer. Hugging Face, one affected company, used a Chinese open-weight model to assess and defend against the attack. In October, a raft of websites went offline after an AI tool found serious flaws in foundational web technology.

Rafe Pilling, senior director of threat intelligence at Sophos, noted the trade-off: freely available models can be downloaded and customized by threat actors. "It's clearly a dual-use technology," he said. He added that any advanced model could go rogue without careful supervision, regardless of origin, and that monitoring and controls matter more than nationalistic narratives.

Source: Wired