Skip to content
Leaderboard / Google

Gemini 3.1 Pro Preview

Google CLOSED Proprietary
Verdict

Gemini 3.1 Pro Preview is a solid mid-pack model: #86 of 406 by SIscope score. Strongest at chat quality, weaker at agents & tools. On the pricey side: costs more than 79% of priced models.

SIscope score
70
#86 of 406
Human-vote rank
#20
of 419 · score 1480
Price per 1M tokens, blended
$4.5
$2 in · $12 out
Context window
1M
1,048,576 tokens

Shape

Each spoke is a category, 0 at the centre, 100 at the edge.

CodingAgentsReasoningMathChatInstructions

Category scores

Percentiles: 90 means better than about 90% of tested models.

Coding68
Agents & tools30
Reasoning & knowledge95
Math81
Chat quality96
Following instructions92

What goes into each category · Compare with DeepSeek V4.1 Flash

Human-vote score over time

Text arena, recent updates.

147514801485Oct 15: 1480.2 (rank #19)Oct 20: 1480.1 (rank #19)Oct 22: 1480.2 (rank #19)Oct 28: 1480.1 (rank #20)Oct 15Oct 28

Coding votes over time

WebDev arena, recent updates.

144014481455Oct 16: 1447 (rank #73)Oct 18: 1446.9 (rank #74)Oct 19: 1446.9 (rank #76)Oct 22: 1446.8 (rank #77)Oct 23: 1446.7 (rank #78)Oct 27: 1446.6 (rank #79)Oct 28: 1446.6 (rank #80)Oct 16Oct 28

Independent test runs

Run by Epoch AI on the same harness for every model. Bars show the share of questions answered correctly.

GPQA Diamond94%
PhD-level science multiple choice questions
Better than 97% of 333 tested models. Run Feb 20, 2026.
OTIS Mock AIME 2024-2596%
Competition maths problems in the style of AIME
Better than 85% of 313 tested models. Run Feb 20, 2026.
FrontierMath Tiers 1-360%
Very hard, unpublished research-style maths problems
Better than 66% of 122 tested models. Run Jun 11, 2026.

Human-vote arenas

People compare two anonymous answers and pick the better one. Scores work like chess ratings.

ArenaRankScoreRange (95%)VotesUpdated
Text arena
Everyday chat, blind human votes
#20 of 41914801477–1483132,547Oct 28, 2026
WebDev arena
Building web apps, blind human votes
#80 of 15214471442–145224,129Oct 28, 2026
Vision arena
Questions about images, blind human votes
#22 of 16712961291–130147,114Oct 19, 2026
Agent arena
Real multi-step agent sessions
#40 of 57-4.6%-6.0–-3.2%88,001Nov 2, 2026

Source: Arena (formerly LMArena) leaderboard dataset, CC BY 4.0. Fetched Nov 4, 2026.

Price

Input
$2
per 1M tokens
Output
$12
per 1M tokens
Blended
$4.5
3 in : 1 out, per 1M
Context window
1M
max tokens per request
Speed
Coming soon
our own tests

List price from the OpenRouter models API (google/gemini-3.1-pro-preview), Nov 4, 2026. Providers can charge differently.