Compare / Claude Fable 5.1 vs gpt-oss-120b
Claude Fable 5.1 vs gpt-oss-120b
21 results from 5 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | Claude Fable 5.1 | gpt-oss-120b |
|---|---|---|---|
| APEX-Agentsreported | Tasks passed% of tasks | 68.6% | 4.4% |
| Arena (formerly LMArena)reported | OverallArena rating | 1,501 | 1,385 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,485 | 1,351 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,484 | 1,277 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,516 | 1,357 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,497 | 1,323 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,501 | 1,352 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,496 | 1,309 |
| Artificial Analysisreported | Artificial Analysis Coding Indexindex score | 81.6 | 30.4 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 53.4 | 11.6 |
| Artificial Analysisreported | GPQA Diamond% of questions | 93.7% | 78.2% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 59.1% | 19.6% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 85.3% | 52.0% |
| Artificial Analysisreported | SciCode% of problems | 63.1% | 34.0% |
| Artificial Analysisreported | Terminal-Bench 2.1% of tasks | 91.4% | 26.2% |
| Artificial Analysisreported | Terminal-Bench 4.0% of tasks | 55.1% | 0.0% |
| Artificial Analysisreported | Τ-bench banking% of tasks | 47.2% | 12.8% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 3.0% | 107.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.43 | 0.29 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 63.9 | 38.5 |
| Vending-Bench 2reported | Money after a yearUS dollars | $5,421.56 | $-21.53 |
Bold green marks the better value on that metric. Where a source reports ranges that overlap, the difference may not be meaningful; see the benchmark page for ranges.