Atlas

Models

← All models

o3

OpenAI · 2025-04-16

o3 is OpenAI's o-series reasoning model that uses extended chain-of-thought to tackle complex problems in math, science, coding, and visual reasoning. It was OpenAI's flagship reasoning model and supports tool use within its reasoning process.

Benchmark scores

BenchmarkScore
AA-LCR69.3
AA-Omniscience Index-15.3
Aider Polyglot81.3
AIME 2024 (avg@8)87.2
AIME 2024-202585.3
AIME 202589.2
AIME 2025 (avg@8)83.3
ALE-Bench1304
APEX (Mercor)8.6
APEX-Agents17.2
APEX-v1-extended60.5
ARC-AGI-141.5
ARC-AGI-23.0
Artificial Analysis Intelligence Index30.4
Artificial Analysis Omniscience Accuracy38.4
Artificial Analysis Omniscience Hallucination Rate87.1
Artificial Analysis Openness Index5.6
BAZAAR TrueSkill Mu7.4
BioTIER Refusal75.6
BRUMO 202595.8
BullshitBench v2 Leaderboard26.0
BuseyBench SVG4.6
CadEval74.0
CadQueryEval36.0
CadQueryEval — Bounding Box44.0
CadQueryEval — Chamfer Distance48.0
CadQueryEval — Code Execution64.0
CadQueryEval — Component Count60.0
CadQueryEval — Hausdorff 95p40.0
CadQueryEval — STL Creation64.0
Loading Atlas data…