Claude Fable 5 vs GPT-5.4 Pro
Claude Fable 5 
GPT-5.4 Pro Verdict
- Claude Fable 5 leads on quality: 70.4 vs 64.2.
- Claude Fable 5 is stronger in agents & tools, coding, composite, human preference, long context, reasoning.
- GPT-5.4 Pro is stronger in instruction following, knowledge, multimodal.
- Claude Fable 5 is 3.4× cheaper ($20.00 vs $67.50 per 1M blended).
- Claude Fable 5 streams 62.9× faster (63 vs 1 tokens per second).
| Metric | Claude Fable 5 | GPT-5.4 Pro |
|---|---|---|
| BenchLeader Index | 70.4 | 64.2 |
| Agents & tools score | 74.4 | – |
| Coding score | 70.6 | – |
| Composite score | 91.6 | – |
| Human preference score | 71.3 | – |
| Instruction following score | 62.6 | 67.3 |
| Knowledge score | 71.3 | 73.2 |
| Long context score | 67.8 | – |
| Multimodal score | 69.6 | 69.8 |
| Reasoning score | 75.6 | 75.1 |
| Blended price $/M | $20.00 | $67.50 |
| Output speed | 63 tok/s | 1 tok/s |
| Time to first answer | 92.2 s | 6.7 s |
| Context window | 1M | 1.1M |
| Humanity's Last Exam | – | 44.3% |
| SimpleBench | – | 74.1% |
| Remote Labor Index | 16.1% | – |
| APEX-Agents | 45.0% | – |
| FrontierCode | 53.5% | – |
| Epoch Capabilities Index | 163.4 | 158.9 |
| LMArena Text | 1507 | – |
| LMArena Hard Prompts | 1533 | – |
| LMArena Coding | 1551 | – |
| LMArena WebDev | 1628 | – |
| LMArena Vision | 1330 | – |
| AA Intelligence Index | 49.7 | – |
| IFBench | 63.5% | – |
| AA-LCR | 82.3% | – |
| AA-Omniscience | 43.3 | – |
| Terminal-Bench Hard | 62.9% | – |
| GPQA Diamond (AA) | 92.6% | – |
| Humanity's Last Exam (AA) | 55.5% | – |
| SciCode (AA) | 61.0% | – |
| τ²-Bench Telecom (AA) | 98.5% | – |
| LiveCodeBench | 89.8% | – |
| MMLU-Pro | 91.5% | – |
| IOI | 72.3% | – |
| LegalBench | 88.6% | – |
| CorpFin | 71.8% | – |
| TaxEval | 76.9% | – |
| Terminal-Bench 2.1 (Vals) | 80.5% | – |
| SWE-bench (Vals) | 95.0% | – |
| GPQA Diamond (Vals) | 93.2% | – |
| Vals Index | 66.0 | – |
| MCP Atlas | 83.3% | – |
| MultiChallenge | – | 69.2% |
| PRBench Finance | 53.9% | – |
| PRBench Legal | 52.6% | – |
| VISTA | – | 53.9% |
| MultiNRC | – | 62.3% |
| HiL-Bench | 56.3% | – |
| TutorBench | – | 56.6% |
| EQ-Bench 4 | 1340 | – |
| Kagi LLM Benchmark | 88.8% | – |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.