Grok 4.5 vs Muse Spark 1.2
Grok 4.5 
Muse Spark 1.2 Verdict
- Grok 4.5 and Muse Spark 1.2 are level on quality (63.4 vs 62.6).
- Grok 4.5 is stronger in agents & tools, human preference, long context, multimodal.
- Muse Spark 1.2 is stronger in coding, composite, knowledge, reasoning, maths.
- Muse Spark 1.2 is 1.5× cheaper ($2.00 vs $3.00 per 1M blended).
- Muse Spark 1.2 streams 3.9× faster (219 vs 56 tokens per second).
| Metric | Grok 4.5 | Muse Spark 1.2 |
|---|---|---|
| BenchLeader Index | 63.4 | 62.6 |
| Agents & tools score | 58.6 | – |
| Coding score | 59.9 | 66.5 |
| Composite score | 66.0 | 79.1 |
| Human preference score | 66.9 | – |
| Knowledge score | 76.0 | 76.9 |
| Long context score | 66.2 | 66.0 |
| Multimodal score | 64.8 | – |
| Reasoning score | 68.9 | 70.2 |
| Maths score | – | 54.3 |
| Blended price $/M | $3.00 | $2.00 |
| Output speed | 56 tok/s | 219 tok/s |
| Time to first answer | 12.7 s | 23.7 s |
| Context window | 500k | 1.0M |
| SimpleBench | 70.0% | 74.5% |
| WeirdML | 46.4% | – |
| APEX-Agents | 34.2% | – |
| FrontierCode | 42.4% | – |
| ProofBench | – | 43.0% |
| Epoch Capabilities Index | 153.9 | 155.5 |
| LMArena Text | 1471 | – |
| LMArena Hard Prompts | 1495 | – |
| LMArena Coding | 1523 | – |
| LMArena WebDev | 1556 | – |
| LMArena Vision | 1291 | – |
| LMArena Agent | 3.9 | – |
| LiveBench | 75.8% | – |
| LiveBench Reasoning | 87.2% | – |
| LiveBench Coding | 68.6% | – |
| LiveBench Agentic Coding | 56.5% | – |
| LiveBench Mathematics | 90.8% | – |
| LiveBench Data Analysis | 73.0% | – |
| LiveBench Language | 82.8% | – |
| AA Intelligence Index | 39.1 | 39.8 |
| AA-LCR | 79.3% | 79.0% |
| MMMU-Pro | 80.4% | – |
| AA-Omniscience | 25.3 | 27.2 |
| GPQA Diamond (AA) | 93.1% | 90.4% |
| Humanity's Last Exam (AA) | 42.7% | 45.5% |
| SciCode (AA) | 55.0% | 57.4% |
| IOI | – | 49.5% |
| Kagi LLM Benchmark | 83.5% | – |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.