Up to four models
GPT-5 vs gpt-oss-120b
On the index (GPT-4 = 100), GPT-5 scores 149 and gpt-oss-120b 129. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
GPT-5
149133–164
WeightsClosed
ReleasedAug 2025
Price in / out$1.25 / $10
Context400K
Core tests taken6 of 9
gpt-oss-120b
129113–145
WeightsOpen
ReleasedAug 2025
Price in / out$0.15 / $0.60
Context131K
Core tests taken3 of 9
GPT-5 leads gpt-oss-120b by 19 points, ahead on 2 of 2 shared core tests, at 13× the price.
Test by testthird-party results
GPT-5gpt-oss-120b
Every result as a table
| Test | GPT-5 | gpt-oss-120b |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 86.2% | 75.8% |
| GPQA Diamond (Vals AI) | 85.6% | 78.5% |
| MMLU Pro | 86.5% | 79.2% |
| Fiction.LiveBench | 97.2% | 44.4% |
| SimpleBench | 56.7% | 22.1% |
| Maths | ||
| AIME-style maths (OTIS mock) | 91.4% | 88.9% |
| MATH 500 | 96.0% | 94.8% |
| MGSM | 92.8% | 92.0% |
| Coding | ||
| Aider Polyglot | 88.0% | 41.8% |
| SWE-bench Pro (public set) | 41.8% | 16.2% |
| Terminal-Bench 2.0 (Terminus 2 runs) | 35.2% | 18.7% |
| LiveCodeBench | 85.9% | 83.2% |
| SWE-bench | 69.0% | 33.6% |
| WeirdML | 60.7% | 48.2% |
| Agents | ||
| DTBench | 90.7% | 76.3% |
| LMCA | 40.0% | 22.1% |
| Mystery game puzzles | 23.0% | 2.0% |
| Professional work | ||
| CorpFin v2 | 61.1% | 58.2% |
| LegalBench | 86.0% | 75.9% |
| TaxEval v2 | 73.4% | 71.6% |
Task length · METR42 sec34 sec
Tests at least two of these models have taken. Bold rows are the core tests.