Compare · ModelsLive · 2 picked · head to head
Claude Opus 4.8 vs Gemini 3 Pro
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Claude Opus 4.8 wins on 13/18 benchmarks
Claude Opus 4.8 wins 13 of 18 shared benchmarks. Leads in agentic · reasoning · arena.
Category leads
agentic·Claude Opus 4.8reasoning·Claude Opus 4.8arena·Claude Opus 4.8knowledge·Claude Opus 4.8math·Claude Opus 4.8coding·Claude Opus 4.8general·Claude Opus 4.8
Hype vs Reality
Attention vs performance
Claude Opus 4.8
#75 by perf·#9 by attention
Gemini 3 Pro
#71 by perf·#5 by attention
Vendor risk
Who is behind the model
Anthropic
$965.0B·Tier 1
Google DeepMind
$4.20T·Tier 1
Head to head
18 benchmarks · 2 models
Claude Opus 4.8Gemini 3 Pro
APEX-Agents
Claude Opus 4.8 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 4.8
48.9
Gemini 3 Pro
18.4
ARC-AGI
Claude Opus 4.8 leads by +17.5
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
Claude Opus 4.8
92.5
Gemini 3 Pro
75.0
ARC-AGI-2
Claude Opus 4.8 leads by +41.0
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 4.8
72.1
Gemini 3 Pro
31.1
Chatbot Arena Elo · Coding
Claude Opus 4.8 leads by +95.1
Claude Opus 4.8
1534.1
Gemini 3 Pro
1439.0
Chatbot Arena Elo · Overall
Gemini 3 Pro leads by +10.9
Claude Opus 4.8
1474.5
Gemini 3 Pro
1485.5
Chess Puzzles
Claude Opus 4.8 leads by +3.2
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
Claude Opus 4.8
30.6
Gemini 3 Pro
27.4
DeepResearch Bench
Claude Opus 4.8 leads by +3.9
DeepResearch Bench · evaluates AI on complex multi-step research tasks requiring information gathering, synthesis, and producing comprehensive analyses.
Claude Opus 4.8
50.2
Gemini 3 Pro
46.3
FrontierMath-2025-02-28-Private
Gemini 3 Pro leads by +18.7
FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning.
Claude Opus 4.8
47.2
Gemini 3 Pro
66.0
FrontierMath-Tier-4-2025-07-01-Private
FrontierMath Tier 4 (Jul 2025) · the most challenging tier of frontier mathematics, containing problems that push the absolute limits of AI mathematical reasoning.
Claude Opus 4.8
31.3
Gemini 3 Pro
31.3
GPQA diamond
Gemini 3 Pro leads by +2.1
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Claude Opus 4.8
88.0
Gemini 3 Pro
90.2
GSO-Bench
Claude Opus 4.8 leads by +28.4
GSO-Bench · evaluates AI models on real-world open-source software engineering tasks, testing the ability to understand and resolve actual GitHub issues.
Claude Opus 4.8
47.1
Gemini 3 Pro
18.6
OTIS Mock AIME 2024-2025
Claude Opus 4.8 leads by +7.0
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Claude Opus 4.8
98.3
Gemini 3 Pro
91.4
PostTrainBench
Claude Opus 4.8 leads by +15.7
Claude Opus 4.8
33.8
Gemini 3 Pro
18.1
Proofbench
Claude Opus 4.8 leads by +49.0
Claude Opus 4.8
69.0
Gemini 3 Pro
20.0
Remote Labor Index
Claude Opus 4.8 leads by +7.1
Claude Opus 4.8
8.3
Gemini 3 Pro
1.3
SimpleBench
Gemini 3 Pro leads by +13.9
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
Claude Opus 4.8
57.8
Gemini 3 Pro
71.7
SimpleQA Verified
Gemini 3 Pro leads by +19.9
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
Claude Opus 4.8
53.0
Gemini 3 Pro
72.9
WeirdML
Claude Opus 4.8 leads by +13.0
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
Claude Opus 4.8
82.9
Gemini 3 Pro
69.9
Full benchmark table
| Benchmark | Claude Opus 4.8 | Gemini 3 Pro |
|---|---|---|
APEX-Agents APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments. | 48.9 | 18.4 |
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 92.5 | 75.0 |
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. | 72.1 | 31.1 |
Chatbot Arena Elo · Coding | 1534.1 | 1439.0 |
Chatbot Arena Elo · Overall | 1474.5 | 1485.5 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 30.6 | 27.4 |
DeepResearch Bench DeepResearch Bench · evaluates AI on complex multi-step research tasks requiring information gathering, synthesis, and producing comprehensive analyses. | 50.2 | 46.3 |
FrontierMath-2025-02-28-Private FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning. | 47.2 | 66.0 |
FrontierMath-Tier-4-2025-07-01-Private FrontierMath Tier 4 (Jul 2025) · the most challenging tier of frontier mathematics, containing problems that push the absolute limits of AI mathematical reasoning. | 31.3 | 31.3 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 88.0 | 90.2 |
GSO-Bench GSO-Bench · evaluates AI models on real-world open-source software engineering tasks, testing the ability to understand and resolve actual GitHub issues. | 47.1 | 18.6 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 98.3 | 91.4 |
PostTrainBench | 33.8 | 18.1 |
Proofbench | 69.0 | 20.0 |
Remote Labor Index | 8.3 | 1.3 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 57.8 | 71.7 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 53.0 | 72.9 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 82.9 | 69.9 |
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 | |
| — | — | — | — |