# Mistral says its open-weight model Le Chonk leads China's models in coding and cybersecurity

published: 2026-10-30T00:05:25.843Z
topic: infrastructure
source: https://www.wired.com/story/mistral-says-le-chonk-surpasses-chinese-models-cybersecurity/
url: https://aiagentnews.io/stories/mistral-says-its-open-weight-model-le-chonk-leads-china-s-models-in-coding-and

Mistral says the finalized version of its trillion-parameter open-weight model, Le Chonk, matches or surpasses models from Chinese labs DeepSeek and Z.ai in automated coding and cybersecurity. The model is now freely available for anyone to access.

French company Mistral released an early version of Le Chonk a few weeks earlier, when it said the model was the most advanced open-weight offering outside China. After further refinement, the company now calls it the best open-weight model for automated coding and cybersecurity. Mistral says it improved the model by training it on increasingly difficult cybersecurity and coding problems.

Pierre Stock, Mistral's VP of science, told reporters that the model offers frontier capabilities for enterprises and governments facing a growing cyber threat. Mistral argues that Western organizations need their own advanced open-weight models, because they cannot be sure Chinese models will not behave maliciously.

Security researchers note the trade-off: freely downloadable models can also be modified and misused by threat actors. Rafe Pilling of Sophos says the risks of advanced models are not specific to any one country and call for monitoring and controls.

## Why it matters if you build agents

Developers building agents now have another freely available open-weight model to evaluate for coding and security tasks, and Mistral says it can be run without a provider able to switch it off remotely.

Because the weights are open, teams that adopt the model will need their own monitoring and access controls, since the source notes that any advanced model can behave unexpectedly.