Compare / Claude Fable 5 vs GPT-5.4 Nano
Claude Fable 5 vs GPT-5.4 Nano
20 results from 3 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | Claude Fable 5 | GPT-5.4 Nano |
|---|---|---|---|
| Arena (formerly LMArena)reported | OverallArena rating | 1,483 | 1,418 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,503 | 1,405 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,501 | 1,335 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,550 | 1,438 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,509 | 1,387 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,504 | 1,401 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,508 | 1,362 |
| Artificial Analysisreported | Artificial Analysis Coding Indexindex score | 76.5 | 56.1 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 49.6 | 20.7 |
| Artificial Analysisreported | GPQA Diamond% of questions | 92.6% | 81.7% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 55.5% | 28.3% |
| Artificial Analysisreported | IFBench% of instructions | 63.5% | 75.9% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 82.3% | 76.7% |
| Artificial Analysisreported | SciCode% of problems | 61.0% | 47.2% |
| Artificial Analysisreported | Terminal-Bench 2.1% of tasks | 84.6% | 60.7% |
| Artificial Analysisreported | Terminal-Bench 4.0% of tasks | 42.4% | 0.5% |
| Artificial Analysisreported | Terminal-Bench Hard% of tasks | 62.9% | 42.4% |
| Artificial Analysisreported | Τ-bench banking% of tasks | 38.1% | 27.4% |
| Artificial Analysisreported | Τ²-bench telecom% of tasks | 98.5% | 76.0% |
| SimpleQA Verified (Epoch AI)reported | Correct answers% of questions | 70.7% | 11.7% |
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.