o3 (medium)
ModelActiveby OpenAI · family “o3-medium”
Epoch Capabilities Index
Not scored by Epoch AI
Released
16 Apr 2025
Context window
Unknown
Max output
Unknown
Input $ / 1M tokens
Unknown
Output $ / 1M tokens
Unknown
Cached input $ / 1M
Unknown
Each benchmark shown separately with its own source
Benchmark results (20)
| Benchmark | Domain | Score | vs best recorded | Setting | Run | Source |
|---|---|---|---|---|---|---|
| Mystery Game Puzzles | games | 23.0% ±4.2 | 27% | medium | 27 Aug 2026 | Epoch ↗ |
| FrontierMath-Tiers-1-3-v2-Private | math | 29.8% ±2.7 | 32% | medium | 27 Aug 2026 | Eval log ↗ |
| Chess Puzzles | games | 38.0% ±4.9 | 53% | medium | 7 Aug 2026 | Eval log ↗ |
| GDPval | agents | 30.8% | 62% | medium | — | External ↗ |
| DeepResearch Bench | agents | 45.2% | 82% | medium | — | External ↗ |
| CadEval | coding | 74.0% | 100% | medium | — | External ↗ |
| ARC-AGI-2 | reasoning | 3.0% | 3% | medium | — | External ↗ |
| METR Time Horizons | agents | 65.4% | 77% | medium | — | External ↗ |
| GeoBench | multimodal | 74.0% | 84% | medium | — | External ↗ |
| FrontierMath-2025-02-28-Privatesuperseded | math | 16.9% ±2.2 | 32% | medium | 16 Nov 2025 | Epoch ↗ |
| Fiction.LiveBench | long-context | 88.9% | 91% | medium | — | External ↗ |
| Lech Mazur Writing | other | 83.9% | 98% | medium | — | External ↗ |
| VPCT | multimodal | 52.0% | 57% | medium | — | External ↗ |
| HLE | knowledge | 19.2% | 35% | medium | — | External ↗ |
| Aider polyglot | coding | 76.9% | 87% | medium | — | External ↗ |
| OTIS Mock AIME 2024-2025 | math | 84.4% ±5.5 | 84% | medium | 7 Aug 2026 | Eval log ↗ |
| SWE-Bench verified | coding | 62.3% ±2.2 | 75% | medium | 12 Feb 2026 | Eval log ↗ |
| OSWorld | agents | 23.0% | 32% | agent: o3 (100 steps) | — | External ↗ |
| GPQA diamond | science | 80.8% ±2.8 | 84% | medium | 7 Aug 2026 | Eval log ↗ |
| ARC-AGI | reasoning | 53.8% | 55% | medium | — | External ↗ |
Source: Epoch AI Benchmarking Hub (CC BY 4.0). “External” rows are leaderboard results Epoch collects from third parties. Best reported setting per benchmark is shown.
API list price over time
Price history
$ per 1M tokens
This model has no public API price listing
0 recorded prices on — (OpenRouter listing); re-read on every data refresh, most recently 8h ago. Steps show when the price changed.
Events
- BenchmarkEpoch AI evaluates o3 (medium)Epoch AI Benchmarking Hub