Inference
EN
Google

Gemini 3.1 Pro

Gemini 3.1 Pro is a reasoning model from Google. 38 benchmarks count toward its index, in 7 categories.

  • Closed weights
  • Reasoning
  • Input: text
  • Released 2026-02
RankedEnough results for an index.
IndexAll categories, out of 100.62.3±3.1Median 51.0
CoverageIndex weight that has results.95%
SpeedOutput tokens per second.151/sMedian 67/s
First tokenSeconds to the first output token.19.9 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$2.00Median $0.60
Output / 1MUS dollars per 1M output tokens.$12.00Median $2.55
ContextMaximum tokens in one request.1MMedian 262K
EloLMArena rating and rank.—Median 1422

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.9 counted58.479 / 218
CodingCode writing and repair.6 counted58.079 / 209
ReasoningLogic problems and puzzles.6 counted65.144 / 227
MultimodalTasks with images and text.7 counted63.532 / 84
KnowledgeFacts and expert knowledge.4 counted69.422 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted61.1—
MathMath problems.5 counted61.743 / 168

Results

38 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
Artificial Analysis GPQA DiamondKnowledge94.1%5 / 199
GPQA diamond · Epoch AIKnowledge94.4% ±1.69 / 217
Massive Multi-discipline Multimodal Understanding ProMultimodal83.9%3 / 53
GPQA Diamond · BenchLMKnowledge94.3%5 / 84
SimpleQA VerifiedKnowledgeAbove fitFar above this model's index.73.5% ±1.46 / 91
Artificial Analysis IFBenchInstruction77.1%10 / 146
Artificial Analysis Humanity's Last ExamKnowledge47.0%21 / 222
τ²-Bench Tool-Agent-User EvaluationAgentic95.6%18 / 149
Artificial Analysis MMMU-ProMultimodal82.4%14 / 113
Artificial Analysis Long Context ReasoningReasoning82.0%29 / 217
Artificial Analysis SciCodeCoding58.7%16 / 114
Critical Physics TasksReasoning17.7%37 / 214
FrontierMath-2025-02-28-PrivateMath36.9% ±2.813 / 69
OTIS Mock AIME 2024-2025Math95.6% ±3.140 / 200
Chess PuzzlesReasoning51.0% ±5.018 / 90
FrontierMath-Tier-4-2025-07-01-PrivateMath16.7% ±5.412 / 55
ScreenSpot ProMultimodal84.4%5 / 20
Gert Labs Composite Game BenchmarkAgentic56.9%14 / 52
Humanity's Last Exam without toolsKnowledge45.4%15 / 53
Terminal-Bench HardAgentic53.8%7 / 23
Mystery Game PuzzlesReasoning34.0% ±4.825 / 80
MazeBench: visual spatial reasoning in a 3D open worldAgentic1.0%16 / 48
APEX-Agents-AAAgentic32.0%11 / 30
SWE-Bench verified · Epoch AICoding75.6% ±2.012 / 32
SimpleVQAMultimodal72.4%7 / 18
ParseBench: document parsing for AI agents (LlamaIndex leaderboard)Multimodal67.3%17 / 39
ERQAMultimodal69.4%8 / 18
FrontierMath-Tiers-1-3-v2-PrivateMath59.6% ±2.942 / 90
React Native EvalsCoding78.9%8 / 16
ZeroBenchMultimodal29.0%5 / 10
FrontierMath-Tier-4-v2-PrivateMath26.8% ±7.039 / 72
GDPval-AA normalizedAgentic14.6%88 / 133
CharXiv ReasoningMultimodal80.2%34 / 49
EBR-benchReasoning14.3%22 / 31
Claw-EvalAgentic57.8%30 / 42
Artificial Analysis AnalystAgentAgentic41.3%21 / 26
Furniture AssemblyReasoning26.7% ±5.630 / 37
HealthBench HardKnowledge20.6%11 / 13
ScarfBench: Cross-Framework Application Migration in Enterprise JavaCoding37.6%10 / 11
DeepSearchQAAgentic69.7%22 / 24
PostTrainBench v1.1Coding22.0%14 / 15
MirrorCodeCoding8.9% ±4.638 / 39
FrontierMath-2025-02-28-PublicMath88.9% ±11.15 / 64
FrontierMath-Tier-4-2025-07-01-PublicMath0.0%3 / 36
Artificial Analysis Global-MMLU-LiteMultilingual93.2%—
LiveCodeBench ProCoding82.9%—
MedXpertQA MultimodalMultimodal81.3%—
MedXpertQA TextKnowledge71.5%—
Artificial Analysis EnterpriseOps-GymAgentic42.2%—
ResearchClawBenchAgentic13.3%—

Sources

BenchLMCC BY-NC 4.0Epoch AICC BY

History

5 releases
Index64.2 → 62.3
23 Oct 202612 Nov 2026
Output / 1M$12.00 → $12.00
23 Oct 202612 Nov 2026

Questions

How much does Gemini 3.1 Pro cost?
$2.00 per 1M input tokens, $12.00 per 1M output tokens.
How good is Gemini 3.1 Pro?
Its index is 62.3 out of 100. That is place 62 of 283 ranked models.
What is Gemini 3.1 Pro best at?
Knowledge: 69.4 out of 100, place 22 of 268.
How fast is Gemini 3.1 Pro?
151 output tokens per second, 19.9 s to the first token.
What is the context window of Gemini 3.1 Pro?
1M tokens.
Is Gemini 3.1 Pro open source?
No. Its weights are not published.
When was Gemini 3.1 Pro released?
February 2026.

Badge

Gemini 3.1 Pro on Inference360: place 62 of 283, index 62.3
HTML
<a href="https://inference360.ai/models/gemini-3-1-pro"><img src="https://inference360.ai/badges/gemini-3-1-pro.svg" alt="Gemini 3.1 Pro on Inference360: place 62 of 283, index 62.3" height="20"></a>
Markdown
[![Gemini 3.1 Pro on Inference360: place 62 of 283, index 62.3](https://inference360.ai/badges/gemini-3-1-pro.svg)](https://inference360.ai/models/gemini-3-1-pro)

Compare with

Claude Opus 4.762.0GPT-5.3 Codex62.0Qwen3.8-Omni-Flash62.7GLM-5.3-Flash63.1dots3-note Preview63.4

Lower price

dots3-note Preview63.4 · FreeGLM-5.3-Flash63.1 · $0.078/1MLing 3.1 Flash67.2 · $0.135/1M

More from Google

All models
Gemini 3.1 Pro Preview61.5Gemini 3.1 Pro Preview Custom ToolsUnrankedGemini 3 Pro Deep ThinkUnrankedNano Banana 2 (Gemini 3.1 Flash Image Preview)UnrankedGemini 3.1 Flash-Lite51.9