BenchAtlas

Rankings / OpenAI

O3 Pro (2025-06-10)

released 2025-06-10

BenchAtlas Index

as of 2026-07-14
Not enough benchmark coverage for a composite score (needs ≥4 families across ≥3 categories). The per-benchmark evidence below stands on its own — sparse coverage is an evidence gap, not a low score.

Benchmark evidence

27 results
external indices
ECImedium effort147.9 pointsindependentEpoch AI Benchmarking Hub
ECIhigh effort147.9 pointsindependentEpoch AI Benchmarking Hub
ECIlow effort147.9 pointsindependentEpoch AI Benchmarking Hub
reasoning math
ARC-AGI-1older version
split=public_eval · model_type=CoT + Synthesis
high effort63.3 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT + Synthesis
high effort59.3 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT + Synthesis
medium effort58.1 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT + Synthesis
medium effort57.0 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT + Synthesis
low effort50.9 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT + Synthesis
low effort44.3 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT + Synthesis
high effort9.1 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT + Synthesis
high effort7.6 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT + Synthesis
medium effort5.2 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT + Synthesis
high effort4.9 %independentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT + Synthesis
medium effort4.7 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT + Synthesis
high effort4.2 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT + Synthesis
high effort3.9 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT + Synthesis
high effort3.9 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT + Synthesis
medium effort3.5 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT + Synthesis
medium effort3.2 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT + Synthesis
medium effort2.5 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT + Synthesis
low effort2.5 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT + Synthesis
low effort2.2 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT + Synthesis
low effort2.0 %independentARC Prize Leaderboard
ARC-AGI-2
split=public_eval · model_type=CoT + Synthesis
low effort1.9 %independentARC Prize Leaderboard
ARC-AGI-2
split=semi_private · model_type=CoT + Synthesis
medium effort1.9 %independentARC Prize Leaderboard
ARC-AGI-1older version
split=semi_private · model_type=CoT + Synthesis
low effort1.6 usd_per_taskindependentARC Prize Leaderboard
ARC-AGI-1older version
split=public_eval · model_type=CoT + Synthesis
low effort1.5 usd_per_taskindependentARC Prize Leaderboard