OpenAIGPT-4.5
Best configuration ranks #303 of 610 on the BenchLeader Index at 49.6 ±4.9. Last measured 2 Sept 2026. Released 27 Feb 2025.
- Blended price
- –
- Output speed
- –
- First answer
- –
- Context
- 128k
- Overall index50
- Reasoning41
- Coding50
- Agents & tools43
- Maths50
- Human preference64
- Multimodal53
- Long context55
- Composite41
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | default | Source |
|---|---|---|
| GPQA Diamond | 68.7%#159 | Epoch AI Benchmarking Hub |
| Humanity's Last Exam | 5.4%#37 | Scale AI / CAIS |
| SimpleBench | 34.5%#65 | SimpleBench |
| LMArena Hard Prompts | 1440#108 | LMArena |
| ARC-AGI-1 | 10.3%#169 | ARC Prize |
| ARC-AGI-2 | 0.8%#164 | ARC Prize |
Coding
| Benchmark | default | Source |
|---|---|---|
| WeirdML | 39.4%#105 | WeirdML |
| LMArena Coding | 1459#116 | LMArena |
| Aider Polyglot | 44.9%#27 | Aider polyglot leaderboard |
Maths
| Benchmark | default | Source |
|---|---|---|
| OTIS Mock AIME | 37.8%#182 | Epoch AI Benchmarking Hub |
| MATH Level 5 | 78.6%#34 | Epoch AI Benchmarking Hub |
Human preference
| Benchmark | default | Source |
|---|---|---|
| LMArena Text | 1445#73 | LMArena |
Multimodal
| Benchmark | default | Source |
|---|---|---|
| LMArena Vision | 1195#76 | LMArena |
| VISTA | 42.1%#36 | Scale AI SEAL |
| MMMU (validation) | 74.4%#16 | MMMU |
Long context
| Benchmark | default | Source |
|---|---|---|
| Fiction.LiveBench 120k | 63.9%#12 | Fiction.live |
Composite
| Benchmark | default | Source |
|---|---|---|
| Epoch Capabilities Indexnot in index | 136.8#112 | Epoch AI Benchmarking Hub |
| AA Intelligence Index | 9.6#325 | Artificial Analysis |
See also
Data as of 9 Sept 2026. Compare with another model.