Compare · ModelsLive · 2 picked · head to head
Claude Opus 5 vs Gemini 3.1 Pro Preview
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Claude Opus 5 wins on 16/20 benchmarks
Claude Opus 5 wins 16 of 20 shared benchmarks. Leads in agentic · reasoning · coding.
Category leads
agentic·Claude Opus 5reasoning·Claude Opus 5knowledge·Gemini 3.1 Pro Previewcoding·Claude Opus 5general·Claude Opus 5math·Claude Opus 5
Hype vs Reality
Attention vs performance
Claude Opus 5
#21 by perf·#9 by attention
Gemini 3.1 Pro Preview
#137 by perf·#5 by attention
Best value
Gemini 3.1 Pro Preview
1.5x better value than Claude Opus 5
Claude Opus 5
4.6 pts/$
$15.00/M
Gemini 3.1 Pro Preview
6.9 pts/$
$7.00/M
Vendor risk
Who is behind the model
Anthropic
$965.0B·Tier 1
Google DeepMind
$4.20T·Tier 1
Head to head
20 benchmarks · 2 models
Claude Opus 5Gemini 3.1 Pro Preview
APEX-Agents
Claude Opus 5 leads by +30.5
APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.
Claude Opus 5
65.8
Gemini 3.1 Pro Preview
35.3
ARC-AGI
Gemini 3.1 Pro Preview leads by +0.5
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
Claude Opus 5
97.5
Gemini 3.1 Pro Preview
98.0
ARC-AGI-2
Claude Opus 5 leads by +13.3
ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data.
Claude Opus 5
90.4
Gemini 3.1 Pro Preview
77.1
Balrog
Claude Opus 5 leads by +6.4
Balrog · benchmarks AI agents on text-based adventure games, testing language understanding, strategic planning, and long-horizon reasoning.
Claude Opus 5
63.4
Gemini 3.1 Pro Preview
57.0
Chess Puzzles
Gemini 3.1 Pro Preview leads by +13.7
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
Claude Opus 5
39.0
Gemini 3.1 Pro Preview
52.6
Deepswe
Claude Opus 5 leads by +61.9
Claude Opus 5
73.7
Gemini 3.1 Pro Preview
11.7
Dtbench
Claude Opus 5 leads by +1.3
Claude Opus 5
96.5
Gemini 3.1 Pro Preview
95.1
Ebr Bench
Claude Opus 5 leads by +31.4
Claude Opus 5
45.7
Gemini 3.1 Pro Preview
14.3
FrontierMath-Tier-4-v2-Private
Claude Opus 5 leads by +46.3
Claude Opus 5
73.2
Gemini 3.1 Pro Preview
26.8
FrontierMath-Tiers-1-3-v2-Private
Claude Opus 5 leads by +26.0
Claude Opus 5
85.6
Gemini 3.1 Pro Preview
59.6
Furniture Assembly
Claude Opus 5 leads by +44.0
Claude Opus 5
44.0
Gemini 3.1 Pro Preview
0.0
GPQA diamond
Gemini 3.1 Pro Preview leads by +0.8
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Claude Opus 5
91.8
Gemini 3.1 Pro Preview
92.6
Lmca
Claude Opus 5 leads by +12.5
Claude Opus 5
75.8
Gemini 3.1 Pro Preview
63.3
Mystery Game Puzzles
Claude Opus 5 leads by +27.5
Claude Opus 5
54.8
Gemini 3.1 Pro Preview
27.3
OTIS Mock AIME 2024-2025
Claude Opus 5 leads by +3.3
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Claude Opus 5
98.9
Gemini 3.1 Pro Preview
95.6
PostTrainBench
Claude Opus 5 leads by +13.1
Claude Opus 5
35.0
Gemini 3.1 Pro Preview
22.0
Proofbench
Claude Opus 5 leads by +73.0
Claude Opus 5
99.0
Gemini 3.1 Pro Preview
26.0
SimpleBench
Claude Opus 5 leads by +1.2
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
Claude Opus 5
76.7
Gemini 3.1 Pro Preview
75.5
SimpleQA Verified
Gemini 3.1 Pro Preview leads by +13.6
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
Claude Opus 5
59.9
Gemini 3.1 Pro Preview
73.5
WeirdML
Claude Opus 5 leads by +19.7
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
Claude Opus 5
91.8
Gemini 3.1 Pro Preview
72.1
Full benchmark table
| Benchmark | Claude Opus 5 | Gemini 3.1 Pro Preview |
|---|---|---|
APEX-Agents APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments. | 65.8 | 35.3 |
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 97.5 | 98.0 |
ARC-AGI-2 ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data. | 90.4 | 77.1 |
Balrog Balrog · benchmarks AI agents on text-based adventure games, testing language understanding, strategic planning, and long-horizon reasoning. | 63.4 | 57.0 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 39.0 | 52.6 |
Deepswe | 73.7 | 11.7 |
Dtbench | 96.5 | 95.1 |
Ebr Bench | 45.7 | 14.3 |
FrontierMath-Tier-4-v2-Private | 73.2 | 26.8 |
FrontierMath-Tiers-1-3-v2-Private | 85.6 | 59.6 |
Furniture Assembly | 44.0 | 0.0 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 91.8 | 92.6 |
Lmca | 75.8 | 63.3 |
Mystery Game Puzzles | 54.8 | 27.3 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 98.9 | 95.6 |
PostTrainBench | 35.0 | 22.0 |
Proofbench | 99.0 | 26.0 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 76.7 | 75.5 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 59.9 | 73.5 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 91.8 | 72.1 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $5.00 | $25.00 | 1.0M tokens (~500 books) | $100.00 | |
| $2.00 | $12.00 | 1.0M tokens (~524 books) | $45.00 |