GPT-5.4 Pro vs GPT-5.6 Sol
GPT-5.4 Pro Verdict
- GPT-5.6 Sol (high) leads on quality: 68.0 vs 64.2.
- GPT-5.4 Pro is stronger in multimodal, reasoning.
- GPT-5.6 Sol (high) is stronger in instruction following, knowledge, agents & tools, coding, composite, long context.
- GPT-5.6 Sol (high) is 8.4× cheaper ($8.00 vs $67.50 per 1M blended).
- GPT-5.6 Sol (high) streams 59.0× faster (59 vs 1 tokens per second).
| Metric | GPT-5.4 Pro | GPT-5.6 Sol (high) |
|---|---|---|
| BenchLeader Index | 64.2 | 68.0 |
| Instruction following score | 67.3 | 67.6 |
| Knowledge score | 73.2 | 73.7 |
| Multimodal score | 69.8 | 66.5 |
| Reasoning score | 75.1 | 62.9 |
| Agents & tools score | – | 87.6 |
| Coding score | – | 72.0 |
| Composite score | – | 82.5 |
| Long context score | – | 67.4 |
| Blended price $/M | $67.50 | $8.00 |
| Output speed | 1 tok/s | 59 tok/s |
| Time to first answer | 6.7 s | 31.4 s |
| Context window | 1.1M | 1M |
| Humanity's Last Exam | 44.3% | – |
| SimpleBench | 74.1% | – |
| SciCode | – | 56.9% |
| WeirdML | – | 88.8% |
| Epoch Capabilities Index | 158.9 | – |
| AA Intelligence Index | – | 42.5 |
| IFBench | – | 69.2% |
| AA-LCR | – | 81.7% |
| MMMU-Pro | – | 81.8% |
| AA-Omniscience | – | 20.4 |
| Terminal-Bench Hard | – | 62.1% |
| GPQA Diamond (AA) | – | 92.8% |
| Humanity's Last Exam (AA) | – | 46.0% |
| SciCode (AA) | – | 57.8% |
| τ²-Bench Telecom (AA) | – | 83.3% |
| MultiChallenge | 69.2% | – |
| EnigmaEval | – | 37.1% |
| VISTA | 53.9% | – |
| MultiNRC | 62.3% | – |
| TutorBench | 56.6% | – |
| ARC-AGI-1 | – | 97.0% |
| ARC-AGI-2 | – | 85.4% |
| ARC-AGI-3 | – | 2.2% |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.