GPT-5 vs gpt-oss-120b
Head to head on 17 shared benchmarks, with price, context window, and release dates.
Benchmark scores
76evals · sort by any column, ⤢ to expand fullscreen
| Model | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| GPT-5 OpenAI | 88.0% | 48.3% | 61.1% | 1623elo | 72.8 | 68.6% | 6.6% | 45.0% | 54.3% | 0.52 | 1.62 | 82.0% | 37.8% | 74.4% | 73.4% | 12.9% | 0.0% | 46.7% | 91.1% | 94.6% | 0.0% | - | - | 0.15 | 0.0% | 68.1% | 4.58 | - | 57.7% | 80.0% | 80.0% | 91.7% | 32.4% | 0.0% | 88.9% | 83.3% | 100.0% | - | 48.9% | 83.4% | 58.2% | - | - | - | - | - | - | - | 37.5% | 100.0% | 86.1% | 0.0% | 92.7% | 92.7% | 49.6% | 83.7% | 93.3% | - | 10.0% | 0.0% | 65.5% | 60.0% | 0.26 | 0.0% | 18.0% | 43.7% | - | 74.9% | 2.34 | - | 87.5% | 0.86 | 20.1% | 54.8% | 90.0% | 90.0% | 56.7% |
| gpt-oss-120b OpenAI | 41.8% | 66.7% | 58.2% | 959elo | 35.8 | 67.2% | 5.9% | 58.3% | 70.7% | 1.62 | 1.53 | 77.5% | 36.0% | 26.0% | 71.6% | 5.3% | 45.0% | - | - | - | - | 38.9% | 7.6% | - | - | - | - | 13.0% | - | - | - | - | - | - | - | - | - | 57.6 | - | - | - | 51.0% | 60.2% | 38.8% | 50.3% | 48.6% | 68.9% | 39.2% | - | - | - | - | - | - | - | - | - | 80.0% | - | - | - | - | - | - | - | - | 97.9% | - | - | 65.6% | - | - | - | - | - | - | 49.6% |
2 / 2 models
| Attribute | GPT-5 OpenAI | gpt-oss-120b OpenAI |
|---|---|---|
| Description | OpenAI's August 2025 unified frontier model that auto-routes between a fast model and a deeper "thinking" variant. | GPT-Oss 120b is an AI model from OpenAI, released with open weights. |
| Family | GPT 5 | GPT |
| Params | Undisclosed (closed) | - |
| Open weights | closed | open |
| License | Proprietary | apache-2.0 |
| Released | 7 Aug 2025 | 5 Aug 2025 |
| Context | 400,000 | 131,072 |
| Input $/1M | - | $0.15 |
| Output $/1M | - | $0.60 |
| Cache read $/1M | $0.125 | $0.030 |
| Cache write $/1M | - | - |
| Speed | 0 tok/s | 186 tok/s |
| Latency | 0.00 s | 0.50 s |
| Modalities | text, image, file | text |
| Providers | openai | - |
| Publisher | OpenAI | OpenAI |
| Reported scores | GPQA Diamond67.2% (AA) Humanity's Last Exam (HLE)5.9% (AA) LiveCodeBench70.7% (AA) MMLU-Pro77.5% (AA) SciCode36.0% (AA) |
FAQ
- Is GPT-5 or gpt-oss-120b better?
- Across the 17 benchmarks both models report on Sophon, GPT-5 leads on 12 of them. Which one is "better" depends on the benchmark - the table above shows every shared score.
- How do GPT-5 and gpt-oss-120b compare on GPQA Diamond?
- GPT-5 scores 68.6% and gpt-oss-120b scores 67.2% on GPQA Diamond.
- When were GPT-5 and gpt-oss-120b released?
- GPT-5 was released 7 Aug 2025 by OpenAI. gpt-oss-120b was released 5 Aug 2025 by OpenAI.
Comparing something else? Build your own side-by-side.