GPT-5.5 vs Muse Spark 1.3
GPT-5.5 
Muse Spark 1.3 Verdict
- GPT-5.5 leads on quality: 66.6 vs 63.9.
- GPT-5.5 is stronger in agents & tools, human preference, instruction following, long context, reasoning.
- Muse Spark 1.3 (xhigh) is stronger in coding, composite, knowledge, multimodal.
- Muse Spark 1.3 (xhigh) is 5.6× cheaper ($2.00 vs $11.25 per 1M blended).
- Muse Spark 1.3 (xhigh) streams 2.2× faster (193 vs 88 tokens per second).
| Metric | GPT-5.5 | Muse Spark 1.3 (xhigh) |
|---|---|---|
| BenchLeader Index | 66.6 | 63.9 |
| Agents & tools score | 68.4 | 60.4 |
| Coding score | 57.3 | 70.8 |
| Composite score | 77.7 | 78.6 |
| Human preference score | 67.6 | – |
| Instruction following score | 73.3 | – |
| Knowledge score | 73.7 | 75.0 |
| Long context score | 68.8 | 68.1 |
| Multimodal score | 64.9 | 66.6 |
| Reasoning score | 69.7 | – |
| Blended price $/M | $11.25 | $2.00 |
| Output speed | 88 tok/s | 193 tok/s |
| Time to first answer | 62.3 s | 40.0 s |
| Context window | 1.1M | 1M |
| Terminal-Bench | 84.7% | – |
| SimpleBench | 69.0% | – |
| Remote Labor Index | 6.3% | – |
| APEX-Agents | 38.5% | – |
| FrontierCode | 43.0% | – |
| Epoch Capabilities Index | 159.1 | – |
| LMArena Text | 1477 | – |
| LMArena Hard Prompts | 1498 | – |
| LMArena Coding | 1509 | – |
| LMArena WebDev | 1458 | 1625 |
| LMArena Vision | 1296 | – |
| LMArena Agent | 3 | – |
| LiveBench | – | 81.6% |
| LiveBench Reasoning | – | 89.7% |
| LiveBench Coding | – | 81.1% |
| LiveBench Agentic Coding | – | 64.1% |
| LiveBench Mathematics | – | 96.0% |
| LiveBench Data Analysis | – | 79.6% |
| LiveBench Language | – | 82.8% |
| AA Intelligence Index | 38.6 | 45.2 |
| IFBench | 75.8% | – |
| AA-LCR | 84.3% | 83.0% |
| MMMU-Pro | 79.9% | 82.0% |
| AA-Omniscience | 20.5 | 23.1 |
| Terminal-Bench Hard | 60.6% | – |
| GPQA Diamond (AA) | 93.5% | 94.1% |
| Humanity's Last Exam (AA) | 45.8% | 47.5% |
| SciCode (AA) | 55.8% | 59.7% |
| τ²-Bench Telecom (AA) | 93.9% | – |
| Terminal-Bench 2.1 (Vals) | – | 72.3% |
| Vals Index | – | 60.3 |
| HiL-Bench | 39.7% | – |
| EQ-Bench 4 | 1315 | – |
| 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.