GPT-5 Mini Medium
| Released | 2025-08-07 |
|---|---|
| Price per 1M | $0.25 in · $2.00 out |
| Overall score | 48.8 · rank 290 |
| Context window | 400,000 tokens |
| Provider | OpenAI |
| Weights | Proprietary |
| Configuration | Medium |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 48.8 | 290 | 11 |
| Overall: Coding | 48.8 | 277 | 6 |
| Overall: Reasoning | 49.6 | 268 | 8 |
| Overall: Agentic work | 47.9 | 297 | 6 |
| Overall: Frontend | 49.1 | 231 | 2 |
| Coding | 49.5 | 249 | 1 |
| Agents | 47.2 | 224 | 1 |
| Reasoning | 48.0 | 285 | 4 |
| Math | 51.6 | 165 | 4 |
| Multimodal | 47.9 | 124 | 1 |
| Value vs price | −0.3 | 251 |
Evidence
Coding 49.5 ±4.0
| Benchmark | Result | Source | Date |
|---|---|---|---|
| SWE-bench Verified (Epoch run) | 64.7% | Epoch AI | 2026-02-01 |
Agents 47.2 ±2.5
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Terminal-Bench 2.0 | 31.9% | Epoch AI | 2025-11-04 |
Reasoning 48.0 ±2.3
| Benchmark | Result | Source | Date |
|---|---|---|---|
| GPQA Diamond | 71.7% | Epoch AI | 2025-08-07 |
| ARC-AGI-2 | 4.0% | Epoch AI | 2026-11-11 |
| ARC-AGI-1 (ARC Prize verified) | 37.3% | ARC Prize | 2026-11-11 |
| ARC-AGI-1 | 37.3% | Epoch AI | 2026-11-11 |
| ARC-AGI-2 (ARC Prize verified) | 4.0% | ARC Prize | 2026-11-11 |
| Mystery Game Puzzles | 4.0% | Epoch AI | 2026-08-27 |
Math 51.6 ±1.4
| Benchmark | Result | Source | Date |
|---|---|---|---|
| MATH Level 5 | 96.8% | Epoch AI | 2025-08-20 |
| FrontierMath Tiers 1-3 (2025 set) | 20.3% | Epoch AI | 2025-11-13 |
| FrontierMath Tier 4 (2025 set) | 4.2% | Epoch AI | 2025-08-07 |
| OTIS Mock AIME 2024-2025 | 78.3% | Epoch AI | 2025-08-07 |
Multimodal 47.9 ±5.3
Results not used in the score
| Benchmark | Result | Source | Why not used |
|---|---|---|---|
| Terminal-Bench 2.0 | 29.2% | Epoch AI | Lower duplicate run on the same benchmark |
| Terminal-Bench 2.0 | 24.0% | Epoch AI | Lower duplicate run on the same benchmark |
| Terminal-Bench 2.0 | 22.2% | Epoch AI | Lower duplicate run on the same benchmark |
Names matched to this model
gpt-5-mini-medium