Capability gains

Google releases eight Gemma 3 open model checkpoints

Google released Gemma 3 base and instruction-tuned checkpoints at 1B, 4B, 12B and 27B sizes, with multimodal reasoning, long context and broad local and cloud distribution.

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 DOOM64confidence 98/100

Why it moved the index

The family combined improved open-weight capability with multimodality, function calling, long context and deployment from phones to cloud systems, materially increasing diffusion and downstream control difficulty.

AUDIT TRAIL

Assessment history

  1. R1
    Toward 64 · confidence 98

    New March 2025 open-weight family release with eight separately published checkpoint identities and no durable collision.

    12 Aug 2026
SHARE THE FINDINGS

Share this page

DoomBench social sharing card for Google releases eight Gemma 3 open model checkpoints.
  1. DoomBench assesses “Google releases eight Gemma 3 open model checkpoints” as evidence moving toward doom, with magnitude 64 and confidence 98 out of 100 in the capability gains category.

  2. The DoomBench assessment of “Google releases eight Gemma 3 open model checkpoints” is based on reporting from Google and records the editorial rationale, source quality, attribution, and revision history.

  3. DoomBench summarizes “Google releases eight Gemma 3 open model checkpoints” as follows: Google released Gemma 3 base and instruction-tuned checkpoints at 1B, 4B, 12B and 27B sizes, with multimodal reasoning, long context and broad local...

    https://www.doombench.com/news/google-releases-eight-gemma-3-open-model-checkpoints-2025-03-12