GPT-5.4 vs Muse Spark 1.2
GPT-5.4 
Muse Spark 1.2 Verdict
- GPT-5.4 and Muse Spark 1.2 are level on quality (63.4 vs 62.6).
- GPT-5.4 is stronger in agents & tools, human preference, instruction following, long context, multimodal.
- Muse Spark 1.2 is stronger in coding, composite, knowledge, reasoning, maths.
- Muse Spark 1.2 is 2.8× cheaper ($2.00 vs $5.63 per 1M blended).
- Muse Spark 1.2 streams 1.7× faster (219 vs 130 tokens per second).
| Metric | GPT-5.4 | Muse Spark 1.2 |
|---|---|---|
| BenchLeader Index | 63.4 | 62.6 |
| Agents & tools score | 66.1 | – |
| Coding score | 51.6 | 66.5 |
| Composite score | 78.1 | 79.1 |
| Human preference score | 64.5 | – |
| Instruction following score | 71.7 | – |
| Knowledge score | 66.8 | 76.9 |
| Long context score | 67.6 | 66.0 |
| Multimodal score | 64.0 | – |
| Reasoning score | 62.3 | 70.2 |
| Maths score | – | 54.3 |
| Blended price $/M | $5.63 | $2.00 |
| Output speed | 130 tok/s | 219 tok/s |
| Time to first answer | 126.4 s | 23.7 s |
| Context window | 1.1M | 1.0M |
| Terminal-Bench | 81.8% | – |
| SimpleBench | – | 74.5% |
| APEX-Agents | 34.9% | – |
| ProofBench | – | 43.0% |
| Epoch Capabilities Index | 156.9 | 155.5 |
| LMArena Text | 1466 | – |
| LMArena Hard Prompts | 1488 | – |
| LMArena Coding | 1514 | – |
| LMArena WebDev | 1386 | – |
| LMArena Vision | 1293 | – |
| AA Intelligence Index | 39.0 | 39.8 |
| IFBench | 74.0% | – |
| AA-LCR | 82.0% | 79.0% |
| MMMU-Pro | 78.4% | – |
| AA-Omniscience | 5.8 | 27.2 |
| Terminal-Bench Hard | 57.6% | – |
| GPQA Diamond (AA) | 92.0% | 90.4% |
| Humanity's Last Exam (AA) | 43.7% | 45.5% |
| SciCode (AA) | – | 57.4% |
| τ²-Bench Telecom (AA) | 87.1% | – |
| IOI | – | 49.5% |
| HiL-Bench | 9.7% | – |
| EQ-Bench 4 | 1272 | – |
| Kagi LLM Benchmark | 63.8% | – |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.