Mistral Says Its AI Model ‘Le Chonk’ Has Leapfrogged China in the Open-Weight Race

Live updates

1. 1h ago

   Mistral Releases Trillion-Parameter AI Model 'Le Chonk'

   Paris-based artificial intelligence lab Mistral publicly released the weights for its Large 4 model, nicknamed 'Le Chonk'. Trained entirely in Europe, the trillion-parameter model is designed to give companies greater control over their cybersecurity. Mistral asserts that the model leapfrogs Chinese competitors in the open-weight race, matching or outperforming top systems from DeepSeek and Z.ai in coding and cybersecurity. VentureBeat notes it arrives as the top open model outside China with improved coding performance and a 63 percent lower cost-per-task, while ZDNET ranks it alongside Codex GPT-6.1 Terra.

   Why it matters

   The release escalates competition between European developers and established U.S. and Chinese artificial intelligence giants. Open-weight models give enterprise users direct access for downloads, offering enhanced data control particularly for cybersecurity operations. While Mistral claims a leap forward, independent testing offers a mixed outlook on relative performance against Asian competitors.

   What is confirmed

   • Mistral released the weights for its trillion-parameter AI model called Large 4, nicknamed Le Chonk.
   • The model was trained entirely in Europe and targets cybersecurity users.
   • Mistral states that the model matches or outperforms models from DeepSeek and Z.ai in coding and cybersecurity.
   • VentureBeat reports the model has improved coding performance and 63 percent lower cost-per-task.

   Still unconfirmed

   • The-decoder.com reports that Le Chonk still trails China overall, scoring 26 points compared to 32 posted by DeepSeek V4.1 Flash and GLM-5.3.
   • ZDNET claims Le Chonk ranks alongside Codex GPT-6.1 Terra for coding.

   What to watch next

   • Independent benchmark testing to verify performance claims against top models from DeepSeek and Z.ai.
   • Enterprise adoption rates and feedback regarding cybersecurity use cases.
   • Response from U.S. and Chinese artificial intelligence labs to Mistral's pricing and open-weight release.
   Sources used for this update (11)

   1. WIRED — Mistral Says Its AI Model ‘Le Chonk’ Has Leapfrogged China in the Open-Weight Race
   2. Reuters — Mistral AI publicly releases Large 4, targets cybersecurity users
   3. TechCrunch — Mistral releases weights for its trillion-parameter AI model, targeting cybersecurity
   4. ZDNET — Mistral's Le Chonk is now a top open model for coding, ranking alongside Codex GPT-6.1 Terra
   5. Euronews.com — Mistral releases latest AI model for download, giving companies more control over their cybersecurity
   6. VentureBeat — Mistral Large 4 'Le Chonk' arrives as the top open model outside China with improved coding performance and 63% lower cost-per-task
   7. www.wired.com — Mistral Says Its AI Model ‘Le Chonk’ Has Leapfrogged China in ...
   8. the-decoder.com — Paris-based Mistral releases "Le Chonk," a strong coding ...
   9. dnyuz.com — Mistral Says Its AI Model ‘Le Chonk’ Has Leapfrogged China in ...
   10. tech.yahoo.com — Mistral is taking on U.S. and Chinese AI giants with its ...
   11. grafy.ai — Mistral Le Chonk Open-Weight AI — Grafy News — Grafy News
   confidence 90%