VerdictPal
Join

Models / Google

Gemini 3 Pro

Gemini 3 · Released 2025-11-18

Google's November 2025 Pro-tier Gemini 3. LegalBench 87% on Vals. Superseded by 3.1 Pro for most frontier work but still on API.

Prior-generation Pro Gemini for workflows already pinned to gemini-3-pro endpoints.

#24 of 51 on AA Index · current snapshot
40 AA Index · 14 benchmark rows · 2 sourcesPosition is benchmark-specific — not a cross-family or cross-source ranking.
Composite70
Context1M
Input / 1M$2.0
Output / 1M$12
Knowledge cutoff2025-01-31
Output speed0 t/s
TTFT (default API)0.00s
Statuslightweight
Gemini
KnowledgeReasoningMathCodingAgenticLong context
Knowledge
90
Reasoning
91
Math
96
Coding
92
Agentic
87
Long context
71

88average across 6 tested families

Benchmark placements

Where Gemini 3 Pro places on each public benchmark source that publishes a row. Rank counts every model with a latest row in the same test — not a universal quality score.

BenchmarkFamilyRankScoreSourceDate
AIMEMath#1/ 1996Artificial Analysis2026-07-21
LiveCodeBenchCoding#1/ 1392Artificial Analysis2026-07-21
MMLU-ProKnowledge#1/ 2190Artificial Analysis2026-07-21
τ³-Bench BankingAgentic#5/ 6487Artificial Analysis2026-07-21
ARC-AGIReasoning#6/ 1476ARC Prize leaderboard2026-02-19
SciCodeCoding#6/ 6556Artificial Analysis2026-07-21
AA-LCRLong context#20/ 6471Artificial Analysis2026-07-21
GPQA DiamondReasoning#20/ 6991Artificial Analysis2026-07-21
Humanity's Last ExamReasoning#21/ 6537Artificial Analysis2026-07-21
IFBenchReasoning#33/ 5770Artificial Analysis2026-07-21
Artificial Analysis Intelligence IndexReasoning#34/ 6740Artificial Analysis2026-07-21
Terminal-BenchAgentic#50/ 6942Artificial Analysis2026-07-21
AA output speedPerformance#63/ 69Artificial Analysis2026-07-21
AA time to first tokenPerformance#63/ 69Artificial Analysis2026-07-21
Family context

The three highest-scoring carded models in each capability family. Where Gemini 3 Pro shows up, it's highlighted.

KnowledgeGemini 3 Pro · 90
  1. 1
    GPT 5.5
    91
  2. 2
    Claude Opus 4.8
    91
  3. 3
    Claude Sonnet 4.6
    91
ReasoningGemini 3 Pro · 91
  1. 1
    Claude Mythos Preview
    95
  2. 2
    GPT-5.6 Sol
    94
  3. 3
    Kimi K3
    94
MathGemini 3 Pro · 96
  1. 1
    OpenAI o3
    99
  2. 2
    OpenAI o1
    97
  3. 3
    Gemini 3 Pro
    96
CodingGemini 3 Pro · 92
  1. 1
    GPT 5.5
    96
  2. 2
    Claude Opus 4.8
    96
  3. 3
    Claude Opus 4.6
    96
AgenticGemini 3 Pro · 87
  1. 1
    GPT 5.5
    100
  2. 2
    Claude Opus 4.8
    100
  3. 3
    Claude Fable 5
    100
Long contextGemini 3 Pro · 71
  1. 1
    Kimi K3
    75
  2. 2
    GPT-5.1
    75
  3. 3
    GPT-5.6 Sol
    74
Newest receipts

Benchmark rows added to the public ledger for Gemini 3 Pro in the last 120 days. Older rows live in the full table below.

13 recent
BenchmarkFamilyScoreSourceDays ago
AA-LCRLong context70.67%Artificial Analysis2
AA output speedPerformance0 t/sArtificial Analysis2
AA time to first tokenPerformance0.00sArtificial Analysis2
AIMEMath95.67%Artificial Analysis2
Artificial Analysis Intelligence IndexReasoning39.6%Artificial Analysis2
GPQA DiamondReasoning90.8%Artificial Analysis2
Humanity's Last ExamReasoning37.2%Artificial Analysis2
IFBenchReasoning70.41%Artificial Analysis2
LiveCodeBenchCoding91.7%Artificial Analysis2
MMLU-ProKnowledge89.8%Artificial Analysis2
SciCodeCoding56.1%Artificial Analysis2
τ³-Bench BankingAgentic87.13%Artificial Analysis2
Terminal-BenchAgentic41.67%Artificial Analysis2
Family coverage

A high composite that hides a weak family is a trap. These bars surface the families where this model hasn't been publicly tested, and where it leads.

Knowledge90
Reasoning91
Math96
Coding92
Agentic87
Long context71
Price vs. performance
0255075100$0.10$0.30$1$3$10$30input USD / 1M tokens · log scale →← intelligence

Hover or tab any dot for name, score, and input price.

AnthropicOpenAIMistralAlibabaDeepSeekGoogleMoonshotMiniMaxxAIMetaZhipuXiaomiMeituanCohere

Full benchmark ledger

Every catalog benchmark for Gemini 3 Pro. Scores link to the original source; gaps mean no public row exists yet.

14 sourced rows

14 of 32 catalog benchmarks have a sourced row for Gemini 3 Pro.

Knowledge

Reasoning

Math

Coding

Agentic

Long context

Performance

Sources