# Muse Spark vs Qwen3 235B-A22B

> Muse Spark is the stronger model overall, scoring 50.3 to 43.2 on the Noometry Index.

- Canonical page: https://noometry.com/compare/muse-spark-vs-qwen3-235b-a22b
- Last updated: 2026-10-10
- Shared benchmarks: 23

## Summary

- They share 23 benchmarks with published results for both. Muse Spark scores higher in 6 categories and Qwen3 235B-A22B in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Muse Spark leads 34.9 to 15.8.
- The biggest single-benchmark swing is CritPt: 11.3% for Muse Spark and 0% for Qwen3 235B-A22B.
- Qwen3 235B-A22B has downloadable open weights; the other is API-only.

## Snapshot

| | Muse Spark | Qwen3 235B-A22B |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 50.3 | 43.2 |
| Rank | 41 | 83 |
| Context | — | 131K |
| Input $/M | — | $0.70 |
| Output $/M | — | $2.80 |
| Weights | Proprietary | Open |

## Coding

- Muse Spark: 45.6 (#65)
- Qwen3 235B-A22B: 43.8 (#72)

| Benchmark | Muse Spark | Qwen3 235B-A22B |
|---|---|---|
| SciCode | 51.5% | 42.4% |
| LMArena Coding | 1482 | 1445 |
| Aider Polyglot | — | 59.6% |
| WeirdML | — | 41% |

## Agentic & Tool Use

- Muse Spark: —
- Qwen3 235B-A22B: 33.7 (#48)

| Benchmark | Muse Spark | Qwen3 235B-A22B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 52.1% |
| Vending-Bench 2 | — | -11.34 |

## Reasoning

- Muse Spark: 34.9 (#61)
- Qwen3 235B-A22B: 15.8 (#302)

| Benchmark | Muse Spark | Qwen3 235B-A22B |
|---|---|---|
| CritPt | 11.3% | 0% |
| LMArena Hard Prompts | 1475 | 1433 |
| Epoch Capabilities Index | 152.09 | 143.87 |
| ARC-AGI-2 | — | 1.3% |
| SimpleBench | — | 31% |
| Kagi LLM Benchmark | — | 69.4% |
| ARC-AGI-1 | — | 11% |
| Chess Puzzles | — | 12% |
| Mystery Game Puzzles | — | 9% |
| DTBench | — | 80.3% |
| LMCA | — | 29.3% |
| ForecastBench | — | 59.7 |

## Math

- Muse Spark: 46.8 (#57)
- Qwen3 235B-A22B: 48.5 (#51)

| Benchmark | Muse Spark | Qwen3 235B-A22B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.9% | 86.7% |
| LMArena Math | 1455 | 1431 |
| FrontierMath (Feb 2025 set) | 39% | 8.5% |
| FrontierMath Tier 4 (v1) | 14.6% | 0% |
| ProofBench | 17% | — |
| Omni-MATH | — | 71.8% |
| MATH Level 5 | — | 68.9% |

## Knowledge

- Muse Spark: 65.7 (#12)
- Qwen3 235B-A22B: 49.5 (#67)

| Benchmark | Muse Spark | Qwen3 235B-A22B |
|---|---|---|
| GPQA Diamond | 89.8% | 80.1% |
| LMArena Expert | 1458 | 1460 |
| Humanity's Last Exam | 40.6% | — |
| SimpleQA Verified | — | 40.4% |
| MMLU-Pro | — | 84.4% |
| Confabulations | — | 15.6% |
| Vectara Hallucination Rate | — | 9.3% |
| GPQA (HELM) | — | 72.7% |

## Multimodal

- Muse Spark: 42.7 (#23)
- Qwen3 235B-A22B: —

| Benchmark | Muse Spark | Qwen3 235B-A22B |
|---|---|---|
| LMArena Vision | 1306 | — |
| LMArena Document | 1444 | — |

## Multilingual

- Muse Spark: 56.3 (#23)
- Qwen3 235B-A22B: 52.5 (#84)

| Benchmark | Muse Spark | Qwen3 235B-A22B |
|---|---|---|
| LMArena Non-English | 1464 | 1410 |
| LMArena Chinese | 1507 | 1481 |
| LMArena French | 1497 | 1445 |
| LMArena German | 1497 | 1432 |
| LMArena Korean | 1457 | 1391 |
| LMArena Russian | 1467 | 1411 |
| LMArena Spanish | 1471 | 1429 |
| LMArena Japanese | — | 1400 |

## Instruction Following

- Muse Spark: 76.1 (#47)
- Qwen3 235B-A22B: 72.8 (#132)

| Benchmark | Muse Spark | Qwen3 235B-A22B |
|---|---|---|
| LMArena Instruction Following | 1443 | 1408 |
| IFEval | — | 83.5% |

## Long Context

- Muse Spark: 44.4 (#65)
- Qwen3 235B-A22B: 46.1 (#25)

| Benchmark | Muse Spark | Qwen3 235B-A22B |
|---|---|---|
| LMArena Longer Query | 1451 | 1426 |
| Fiction.LiveBench | — | 75% |

## Writing & Preference

- Muse Spark: 66.2 (#37)
- Qwen3 235B-A22B: 59.8 (#103)

| Benchmark | Muse Spark | Qwen3 235B-A22B |
|---|---|---|
| LMArena Text | 1474 | 1419 |
| LMArena Creative Writing | 1460 | 1384 |
| LMArena Multi-Turn | 1477 | 1432 |
| Short-Story Creative Writing | — | 83% |
| EQ-Bench Creative Writing | — | 1366 |
| WildBench | — | 86.6% |

## FAQ

### Is Muse Spark better than Qwen3 235B-A22B?

Muse Spark is the stronger model overall, scoring 50.3 to 43.2 on the Noometry Index.

### Is Muse Spark or Qwen3 235B-A22B better for coding?

Muse Spark scores higher on coding benchmarks: 45.6 versus 43.8 in the Noometry coding category.

### How many benchmarks do Muse Spark and Qwen3 235B-A22B share?

23 benchmarks have published results for both models. Muse Spark has 27 scored results on Noometry and Qwen3 235B-A22B has 49.