Compare · ModelsLive · 2 picked · head to head
o1 vs o3
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
o3 wins on 12/13 benchmarks
o3 wins 12 of 13 shared benchmarks. Leads in coding · reasoning · knowledge.
Category leads
coding·o3reasoning·o3knowledge·o3math·o3
Hype vs Reality
Attention vs performance
o1
#57 by perf·no signal
o3
#67 by perf·no signal
Vendor risk
Who is behind the model
OpenAI
$840.0B·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
13 benchmarks · 2 models
o1o3
Aider polyglot
o3 leads by +19.6
Aider Polyglot · measures how well AI models can edit code across multiple programming languages using the Aider coding assistant framework.
o1
61.7
o3
81.3
ARC-AGI
o3 leads by +30.1
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
o1
30.7
o3
60.8
CadEval
o3 leads by +18.0
CadEval · evaluates the ability to generate and reason about Computer-Aided Design code, testing spatial reasoning and engineering knowledge.
o1
56.0
o3
74.0
Fiction.LiveBench
o3 leads by +5.6
Fiction.LiveBench · a continuously updated benchmark using recently published fiction to test reading comprehension and reasoning, preventing data contamination.
o1
83.3
o3
88.9
FrontierMath-2025-02-28-Private
o3 leads by +9.4
FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning.
o1
9.3
o3
18.7
GeoBench
o1 leads by +6.0
GeoBench · tests geographic knowledge and spatial reasoning across countries, landmarks, coordinates, and geopolitical understanding.
o1
80.0
o3
74.0
GPQA diamond
o3 leads by +6.7
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
o1
69.0
o3
75.8
Lech Mazur Writing
o3 leads by +13.7
Lech Mazur Writing · evaluates creative writing ability, assessing prose quality, narrative coherence, and stylistic sophistication.
o1
70.2
o3
83.9
MATH level 5
o3 leads by +3.1
MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics.
o1
94.7
o3
97.8
OTIS Mock AIME 2024-2025
o3 leads by +10.6
OTIS Mock AIME 2024–2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
o1
73.3
o3
83.9
SimpleBench
o3 leads by +15.6
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
o1
28.1
o3
43.7
VPCT
o3 leads by +22.5
VPCT (Visual Pattern Completion Test) · tests visual reasoning and pattern recognition by having models complete visual sequences and transformations.
o1
5.5
o3
28.0
WeirdML
o3 leads by +8.6
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
o1
43.8
o3
52.4
Full benchmark table
| Benchmark | o1 | o3 |
|---|---|---|
Aider polyglot Aider Polyglot · measures how well AI models can edit code across multiple programming languages using the Aider coding assistant framework. | 61.7 | 81.3 |
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 30.7 | 60.8 |
CadEval CadEval · evaluates the ability to generate and reason about Computer-Aided Design code, testing spatial reasoning and engineering knowledge. | 56.0 | 74.0 |
Fiction.LiveBench Fiction.LiveBench · a continuously updated benchmark using recently published fiction to test reading comprehension and reasoning, preventing data contamination. | 83.3 | 88.9 |
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. | 9.3 | 18.7 |
GeoBench GeoBench · tests geographic knowledge and spatial reasoning across countries, landmarks, coordinates, and geopolitical understanding. | 80.0 | 74.0 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 69.0 | 75.8 |
Lech Mazur Writing Lech Mazur Writing · evaluates creative writing ability, assessing prose quality, narrative coherence, and stylistic sophistication. | 70.2 | 83.9 |
MATH level 5 MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics. | 94.7 | 97.8 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024–2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 73.3 | 83.9 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 28.1 | 43.7 |
VPCT VPCT (Visual Pattern Completion Test) · tests visual reasoning and pattern recognition by having models complete visual sequences and transformations. | 5.5 | 28.0 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 43.8 | 52.4 |
Pricing · per 1M tokens · projected $/mo at 10M tokens