Compare · ModelsLive · 2 picked · head to head
Claude Opus 4.6 vs o3
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Claude Opus 4.6 wins on 11/12 benchmarks
Claude Opus 4.6 wins 11 of 12 shared benchmarks. Leads in reasoning · math · knowledge.
Category leads
reasoning·Claude Opus 4.6math·Claude Opus 4.6knowledge·Claude Opus 4.6coding·Claude Opus 4.6
Hype vs Reality
Attention vs performance
Claude Opus 4.6
#54 by perf·#4 by attention
o3
#67 by perf·no signal
Best value
o3
2.9x better value than Claude Opus 4.6
Claude Opus 4.6
3.8 pts/$
$15.00/M
o3
11.0 pts/$
$5.00/M
Vendor risk
Who is behind the model
Anthropic
$380.0B·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
12 benchmarks · 2 models
Claude Opus 4.6o3
ARC-AGI
Claude Opus 4.6 leads by +33.2
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
Claude Opus 4.6
94.0
o3
60.8
ARC-AGI-2
Claude Opus 4.6 leads by +62.6
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 4.6
69.2
o3
6.5
FrontierMath-2025-02-28-Private
Claude Opus 4.6 leads by +22.0
FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning.
Claude Opus 4.6
40.7
o3
18.7
FrontierMath-Tier-4-2025-07-01-Private
Claude Opus 4.6 leads by +20.8
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 4.6
22.9
o3
2.1
GPQA diamond
Claude Opus 4.6 leads by +11.6
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Claude Opus 4.6
87.4
o3
75.8
GSO-Bench
Claude Opus 4.6 leads by +24.5
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 4.6
33.3
o3
8.8
HLE
Claude Opus 4.6 leads by +14.8
HLE (Humanity's Last Exam) · crowdsourced expert-level questions designed to be among the hardest possible challenges for AI systems across all domains.
Claude Opus 4.6
31.1
o3
16.3
OTIS Mock AIME 2024-2025
Claude Opus 4.6 leads by +10.6
OTIS Mock AIME 2024–2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Claude Opus 4.6
94.4
o3
83.9
SimpleBench
Claude Opus 4.6 leads by +17.4
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
Claude Opus 4.6
61.1
o3
43.7
SimpleQA Verified
o3 leads by +6.5
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
Claude Opus 4.6
46.5
o3
53.0
SWE-Bench verified
Claude Opus 4.6 leads by +16.4
Claude Opus 4.6
78.7
o3
62.3
WeirdML
Claude Opus 4.6 leads by +25.5
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
Claude Opus 4.6
77.9
o3
52.4
Full benchmark table
| Benchmark | Claude Opus 4.6 | o3 |
|---|---|---|
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 94.0 | 60.8 |
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. | 69.2 | 6.5 |
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. | 40.7 | 18.7 |
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. | 22.9 | 2.1 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 87.4 | 75.8 |
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. | 33.3 | 8.8 |
HLE HLE (Humanity's Last Exam) · crowdsourced expert-level questions designed to be among the hardest possible challenges for AI systems across all domains. | 31.1 | 16.3 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024–2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 94.4 | 83.9 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 61.1 | 43.7 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 46.5 | 53.0 |
SWE-Bench verified | 78.7 | 62.3 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 77.9 | 52.4 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| Claude Opus 4.6 | $5.00 | $25.00 | 1.0M tokens (~500 books) | $100.00 |
| o3 | $2.00 | $8.00 | 200K tokens (~100 books) | $35.00 |