Mistral’s new 1T model aims to leapfrog closed and open rivals

• Introducing Mistral Large 4

  Mistral AIOct 6, 2026

  Update

  Our summaryRead the original

• Mistral’s new 1T model aims to leapfrog closed and open rivals

  TechCrunch AIOct 6, 2026First report

  New

  Our summaryRead the original

• Mistral AI Releases Mistral Large 4 (Le Chonk): A 1.05T Parameter Multimodal MoE Model

  MarkTechPostOct 6, 2026

  Update

  Our summaryRead the original

• Mistral Large 4

  Simon WillisonOct 6, 2026

  Update

  Our summaryRead the original

• Introducing Mistral Large 4: Le chonk

  Simon WillisonOct 6, 2026

  Update

  Our summaryRead the original

1. Revision 5Assessed 6 Oct

   Update

   Older report added; assessment kept from a newer one

   • Event type: analysis_commentary → model_release
   • Indicators to watch: Whether Mistral publishes the promised weights by the end of October 2026, Independent evaluation of model capabilities and benchmark performance → Whether the weights ship as scheduled, The weights' licence and use restrictions, Independent evaluations of model capabilities
   • Evidence limitations: The supplied text contains no model outputs, scoring method, or comparative results, and the benchmark-saturation statement is a commenter’s opinion. → This is a single independent press report; no primary release record, independent capability evaluation, or licence terms are included in the supplied text.
2. Revision 4Assessed 6 Oct

   Update

   Older report added; assessment kept from a newer one

   • Event type: model_release → analysis_commentary
   • Evidence limitations: This is a single analysis source; performance comparisons are reported without independently verifiable benchmark details, and the promised weight release is prospective. → The supplied text contains no model outputs, scoring method, or comparative results, and the benchmark-saturation statement is a commenter’s opinion.
3. Revision 3Assessed 6 Oct

   Update

   Update from Simon Willison

   • Impact: high → medium
   • Status: assessed → updated
   • Indicators to watch: Whether weights are released by the stated deadline, Weight licence and redistribution terms, Independent reproduction of the Artificial Analysis Cyber Index and Cybench results, Safety findings from real-world red-teaming → Whether Mistral publishes the promised weights by the end of October 2026, Independent evaluation of model capabilities and benchmark performance
   • Evidence limitations: The announcement is from the model developer. It cites an independent benchmark but supplies no underlying evaluation report or independent confirmation here; weights, licence terms and further benchmark details are not yet available. → This is a single analysis source; performance comparisons are reported without independently verifiable benchmark details, and the promised weight release is prospective.
4. Revision 2Assessed 6 Oct

   Update

   Older report added; assessment kept from a newer one

   • Indicators to watch: Whether Mistral releases the weights on the stated timetable, The terms governing weight access and use, Independent benchmark results, Independent verification of training-compute claims → Whether weights are released by the stated deadline, Weight licence and redistribution terms, Independent reproduction of the Artificial Analysis Cyber Index and Cybench results, Safety findings from real-world red-teaming
   • Evidence limitations: This is a single press report of Mistral's announcement and statements; benchmarks are pending, and the planned weights release and performance claims are not independently verified. → The announcement is from the model developer. It cites an independent benchmark but supplies no underlying evaluation report or independent confirmation here; weights, licence terms and further benchmark details are not yet available.
5. First assessmentAssessed 6 Oct

   New

   First assessment, from TechCrunch AI

   • Impact: not set → high
   • Status: developing → assessed
   • Confidence: not set → medium
   • Assessment: not set → Mistral Large 4 is a substantial model announcement, but its capabilities remain unverified and its weights are not yet available.
   • Indicators to watch: (none) → Whether Mistral releases the weights on the stated timetable, The terms governing weight access and use, Independent benchmark results, Independent verification of training-compute claims
   • Evidence limitations: not set → This is a single press report of Mistral's announcement and statements; benchmarks are pending, and the planned weights release and performance claims are not independently verified.
   • Representative source: not set → Mistral’s new 1T model aims to leapfrog closed and open rivals (TechCrunch AI)

No maturity ladder applies to this desk.