# MiniMax-M3 vs Muse Spark 1.2

> Muse Spark 1.2 is the stronger model overall, scoring 50.3 to 43.8 on the Noometry Index. MiniMax-M3 costs 3.8× less per token, which makes it the better buy when Muse Spark 1.2's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/minimax-m3-vs-muse-spark-1-2
- Last updated: 2026-10-12
- Shared benchmarks: 25

## Summary

- They share 25 benchmarks with published results for both. MiniMax-M3 scores higher in 1 category and Muse Spark 1.2 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Muse Spark 1.2 leads 51.3 to 30.1.
- The biggest single-benchmark swing is SimpleBench: 45.8% for MiniMax-M3 and 74.5% for Muse Spark 1.2.
- MiniMax-M3 is cheaper at $0.30 / $1.20 per million input/output tokens, against $1.25 / $4.25 for Muse Spark 1.2.
- Muse Spark 1.2 accepts more context: 1.05M tokens versus 1M.
- MiniMax-M3 has downloadable open weights; the other is API-only.

## Snapshot

| | MiniMax-M3 | Muse Spark 1.2 |
|---|---|---|
| Provider | MiniMax | Meta |
| Noometry Index | 43.8 | 50.3 |
| Rank | 85 | 48 |
| Context | 1M | 1.05M |
| Input $/M | $0.30 | $1.25 |
| Output $/M | $1.20 | $4.25 |
| Weights | Open | Proprietary |

## Coding

- MiniMax-M3: 41.8 (#118)
- Muse Spark 1.2: 49.2 (#51)

| Benchmark | MiniMax-M3 | Muse Spark 1.2 |
|---|---|---|
| LMArena WebDev | 1482 | 1533 |
| SciCode | 47.1% | 56.4% |
| LMArena Coding | 1469 | 1495 |
| DeepSWE | — | 54.9% |
| FrontierCode | 14.7% | — |
| FrontierSWE | — | 12% |
| WeirdML | — | 60.3% |
| ALE-Bench | 640.02 | — |

## Agentic & Tool Use

- MiniMax-M3: 22.6 (#130)
- Muse Spark 1.2: 29.4 (#87)

| Benchmark | MiniMax-M3 | Muse Spark 1.2 |
|---|---|---|
| APEX-Agents | 37.7% | 36.4% |
| OSWorld 2.0 | 4.6% | — |
| GBAEval | 0.9% | — |
| GDP.pdf | — | 16% |
| Vending-Bench 2 | 2,158 | — |

## Reasoning

- MiniMax-M3: 30.1 (#87)
- Muse Spark 1.2: 51.3 (#34)

| Benchmark | MiniMax-M3 | Muse Spark 1.2 |
|---|---|---|
| SimpleBench | 45.8% | 74.5% |
| NYT Connections (extended) | 65.1% | 79.2% |
| CritPt | 3.7% | 17.7% |
| LMArena Hard Prompts | 1447 | 1486 |
| DTBench | 78.9% | 94.7% |
| LMCA | 33.7% | 48.4% |
| Epoch Capabilities Index | 146.95 | 154.87 |
| Chess Puzzles | 14% | — |
| Mystery Game Puzzles | 8% | — |
| Surface Evolver Bench | 55% | — |
| ForecastBench | 61.4 | — |

## Math

- MiniMax-M3: 40.0 (#95)
- Muse Spark 1.2: 46.4 (#70)

| Benchmark | MiniMax-M3 | Muse Spark 1.2 |
|---|---|---|
| ProofBench | 18% | 43% |
| LMArena Math | 1429 | 1471 |
| OTIS Mock AIME 2024-2025 | 71.1% | — |

## Knowledge

- MiniMax-M3: 58.4 (#35)
- Muse Spark 1.2: 54.1 (#53)

| Benchmark | MiniMax-M3 | Muse Spark 1.2 |
|---|---|---|
| LMArena Expert | 1461 | 1480 |
| GPQA Diamond | 90.9% | — |
| SimpleQA Verified | — | 60.3% |

## Multimodal

- MiniMax-M3: 40.2 (#51)
- Muse Spark 1.2: 43.4 (#25)

| Benchmark | MiniMax-M3 | Muse Spark 1.2 |
|---|---|---|
| LMArena Vision | 1253 | 1305 |
| LMArena Document | 1435 | — |

## Multilingual

- MiniMax-M3: 53.0 (#75)
- Muse Spark 1.2: 57.1 (#11)

| Benchmark | MiniMax-M3 | Muse Spark 1.2 |
|---|---|---|
| LMArena Non-English | 1420 | 1478 |
| LMArena Chinese | 1463 | 1511 |
| LMArena French | 1447 | 1513 |
| LMArena Russian | 1428 | 1487 |
| LMArena Spanish | 1432 | 1498 |
| LMArena German | 1426 | — |
| LMArena Japanese | 1381 | — |
| LMArena Korean | 1372 | — |

## Instruction Following

- MiniMax-M3: 75.5 (#62)
- Muse Spark 1.2: 76.7 (#36)

| Benchmark | MiniMax-M3 | Muse Spark 1.2 |
|---|---|---|
| LMArena Instruction Following | 1433 | 1461 |

## Long Context

- MiniMax-M3: 44.2 (#72)
- Muse Spark 1.2: 45.2 (#48)

| Benchmark | MiniMax-M3 | Muse Spark 1.2 |
|---|---|---|
| LMArena Longer Query | 1445 | 1475 |

## Writing & Preference

- MiniMax-M3: 62.1 (#83)
- Muse Spark 1.2: 72.3 (#14)

| Benchmark | MiniMax-M3 | Muse Spark 1.2 |
|---|---|---|
| LMArena Text | 1433 | 1482 |
| LMArena Creative Writing | 1404 | 1449 |
| LMArena Multi-Turn | 1442 | 1494 |
| EQ-Bench Creative Writing | — | 1840 |
| EQ-Bench 4 | 1150 | — |

## FAQ

### Is MiniMax-M3 better than Muse Spark 1.2?

Muse Spark 1.2 is the stronger model overall, scoring 50.3 to 43.8 on the Noometry Index. MiniMax-M3 costs 3.8× less per token, which makes it the better buy when Muse Spark 1.2's lead doesn't matter for your workload.

### Which is cheaper, MiniMax-M3 or Muse Spark 1.2?

MiniMax-M3 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; Muse Spark 1.2 lists at $1.25 and $4.25.

### Is MiniMax-M3 or Muse Spark 1.2 better for coding?

Muse Spark 1.2 scores higher on coding benchmarks: 49.2 versus 41.8 in the Noometry coding category.

### Which has the bigger context window?

Muse Spark 1.2 does, with 1.05M tokens against 1M.

### How many benchmarks do MiniMax-M3 and Muse Spark 1.2 share?

25 benchmarks have published results for both models. MiniMax-M3 has 41 scored results on Noometry and Muse Spark 1.2 has 31.