Competitive race

Mistral Large 3 brings frontier-scale multimodal AI to open weights

Mistral AI released the 675-billion-parameter Mistral Large 3 base and instruction checkpoints under Apache 2.0 with API and multi-platform access.

0 comments · 0 votesOpen discussion

Public discussion is readable by everyone. Sign in to comment, reply, or vote.

No comments yet. Start the discussion.

CURRENT ASSESSMENT · REVISION 1
TOWARD DOOM40confidence 86/100

Why it moved the index

Permissive release of a frontier-scale, multimodal, agent-capable model directly broadens uncontrolled access and competitive pressure at the open frontier.

AUDIT TRAIL

Assessment history

  1. R1
    Toward 40 · confidence 86

    New historical open-weight launch record from Mistral AI's dated release.

    11 Aug 2026
SHARE THE FINDINGS

Share this page

DoomBench social sharing card for Mistral Large 3 brings frontier-scale multimodal AI to open weights.
  1. DoomBench assesses “Mistral Large 3 brings frontier-scale multimodal AI to open weights” as evidence moving toward doom, with magnitude 40 and confidence 86 out of 100 in the competitive race category.

  2. The DoomBench assessment of “Mistral Large 3 brings frontier-scale multimodal AI to open weights” is based on reporting from Mistral AI and records the editorial rationale, source quality, attribution, and revision history.

  3. DoomBench summarizes “Mistral Large 3 brings frontier-scale multimodal AI to open weights” as follows: Mistral AI released the 675-billion-parameter Mistral Large 3 base and instruction checkpoints under Apache 2.0 with API and...

    https://www.doombench.com/news/mistral-large-3-brings-frontier-scale-multimodal-ai-to-open-weights-2025-12-02