Google DeepMind · Gemma 3

Gemma 3 4B Base

READER SUMMARY

A pretrained 4B multimodal Gemma 3 checkpoint with long context and local deployment.

DOOM SCORE46.8out of 100model risk profile, not the overall index

CURRENT ASSESSMENT · REVISION 1

Why this model scores 46.8

Multimodality and long context increase capability while base training limits direct agency. Small open weights enable near-universal deployment and difficult downstream control.

Capability42

Autonomy4

Deployment99

Misuse potential39

Control difficulty58

0 comments · 0 votesOpen discussion

No comments yet. Start the discussion.

MODEL-ATTRIBUTED EVIDENCE

News tied to Gemma 3 4B Base

The model score of 46.8 rates this model's risk profile. The overall Doom Index of 63.5 measures the recalibrated complete evidence corpus. These values answer different questions.

NET MODEL-ATTRIBUTED INDEX CONTRIBUTION+0.00

Each article's leave-one-out Doom Index contribution is divided equally among the exact models named on that article. This prevents multi-model evidence from being claimed in full on several model pages. Model risk scores are recalculated from their technical profile and related news, but never feed back into the overall index.

12 Mar 2025CapabilityTOWARD

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.

Full item contribution: +0.03
Gemma 3 4B Base equal share: +0.00

Read assessment →

AUDIT TRAIL

Model score history

R1
Doom Score 46.8
Initial exact profile for the separately released 4B base checkpoint.
12 Aug 2026