Mistral AI launches 1.05T parameter 'Le Chonk' model, challenging closed AI leaders

Mistral AI has released Mistral Large 4, a new 1.05 trillion parameter model, positioning it as a strong contender in the open-weight model space. The model boasts impressive performance on cybersecurity benchmarks, achieving 82% on the AA Cyber Index, though closed models like Claude Opus 5.5 and GPT-6 Astra reportedly refuse to complete this specific task. While Large 4 shows promise in multilingual capabilities and offers competitive pricing, it falls slightly behind Claude Opus 5 in human-evaluated coding tasks. AI

IMPACT Sets a new benchmark for open-weight models, particularly in cybersecurity, and offers a competitive alternative to closed models.

RANK_REASON Frontier-lab model release with system card and performance benchmarks. [lever_c_demoted from frontier_release: ic=1 ai=1.0]

Read on dev.to — LLM tag →

• Artificial Analysis Cyber Index
• Beam
• Claude Opus 5.5
• DeepSeek
• General Language Model
• GPT-6 Astra
• Le Chonk
• Mistral AI
• Mistral Large 4
• Nvidia Grace Blackwell
• Qwen
• Reflection
• Surge AI

• model release
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →