The Open-Source AI Stack
RSS
All models

Models · gemini-3

Gemini 3.1 Pro

Proprietary Google DeepMind · 2026-02-19 · Proprietary

Preview release of the Gemini 3.1 line, three months after Gemini 3 Pro. DeepMind highlighted complex reasoning and code generation that can ship animated SVGs and synthesize entire dashboards from a prompt. Broader rollout to Pro and Ultra subscribers followed May 4 2026.

Architecture

tokens in Embedding vocab not disclosed × N layers Architecture not disclosed (proprietary or undocumented) Output projection tokens out
Schema-generated from data/models.yaml. Every label is auditable against the model's sources.

Specs

Architecture
unknown
Total params
not disclosed
Active params
not disclosed
Context window
1.0M tokens
Attention
unknown
Position encoding
unknown
Post-training
rlhf
OSI-approved
no
Data released
no
Training code
not released

Benchmarks

Each score carries the date it was published; we never infer or interpolate missing scores.

Available quantizations

None. The weights are not distributed, so there are no public quantizations.

Notable innovations

  • · Double the ARC-AGI-2 score of Gemini 3 Pro per Google
  • · Direct SVG and dashboard generation from prompts

Lineage

3.1 line preview, three months after Gemini 3 Pro.

Sources