Rankings / OpenAI
Codex Mini Latest
released 2025-05-23
BenchAtlas Index
as of 2026-07-14
Benchmark evidence
8 results
reasoning math
| ARC-AGI-1older version split=public_eval · model_type=CoT | 33.4 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-1older version split=semi_private · model_type=CoT | 27.3 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=semi_private · model_type=CoT | 1.3 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=semi_private · model_type=CoT | 0.2 usd_per_task | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=public_eval · model_type=CoT | 0.2 usd_per_task | independent | ARC Prize Leaderboard | |
| ARC-AGI-1older version split=semi_private · model_type=CoT | 0.2 usd_per_task | independent | ARC Prize Leaderboard | |
| ARC-AGI-1older version split=public_eval · model_type=CoT | 0.1 usd_per_task | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=public_eval · model_type=CoT | 0.0 % | independent | ARC Prize Leaderboard |
Agent + model results
systems, not bare-model scores
| agent + model Codex CLI + Codex Mini Latest | Terminal-Bench 1.0 | 11.3 % | community | Terminal-Bench Leaderboard |
These scores measure the whole agent system (scaffold, tools, budgets) — they are never merged into the bare model’s numbers.
