# MiniMax-M3 vs Qwen3 Max

> MiniMax-M3 and Qwen3 Max score almost the same on the Noometry Index (43.8 vs 43.7), so choose on price, context window or the category you care about most.

- Canonical page: https://noometry.com/compare/minimax-m3-vs-qwen3-max
- Last updated: 2026-10-10
- Shared benchmarks: 27

## Summary

- They share 27 benchmarks with published results for both. MiniMax-M3 scores higher in 5 categories and Qwen3 Max in 3 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where MiniMax-M3 leads 58.4 to 48.1.
- The biggest single-benchmark swing is NYT Connections (extended): 65.1% for MiniMax-M3 and 30.1% for Qwen3 Max.
- MiniMax-M3 is cheaper at $0.30 / $1.20 per million input/output tokens, against $1.20 / $6 for Qwen3 Max.
- MiniMax-M3 accepts more context: 1M tokens versus 262K.
- MiniMax-M3 has downloadable open weights; the other is API-only.

## Snapshot

| | MiniMax-M3 | Qwen3 Max |
|---|---|---|
| Provider | MiniMax | Alibaba (Qwen) |
| Noometry Index | 43.8 | 43.7 |
| Rank | 85 | 87 |
| Context | 1M | 262K |
| Input $/M | $0.30 | $1.20 |
| Output $/M | $1.20 | $6 |
| Weights | Open | Proprietary |

## Coding

- MiniMax-M3: 41.8 (#118)
- Qwen3 Max: 43.0 (#93)

| Benchmark | MiniMax-M3 | Qwen3 Max |
|---|---|---|
| LMArena Coding | 1469 | 1456 |
| ALE-Bench | 640.02 | 370.45 |
| FrontierCode | 14.7% | — |
| LMArena WebDev | 1482 | — |
| SciCode | 47.1% | — |

## Agentic & Tool Use

- MiniMax-M3: 22.6 (#130)
- Qwen3 Max: —

| Benchmark | MiniMax-M3 | Qwen3 Max |
|---|---|---|
| Vending-Bench 2 | 2,158 | 71.56 |
| APEX-Agents | 37.7% | — |
| OSWorld 2.0 | 4.6% | — |
| GBAEval | 0.9% | — |

## Reasoning

- MiniMax-M3: 30.1 (#87)
- Qwen3 Max: 22.6 (#190)

| Benchmark | MiniMax-M3 | Qwen3 Max |
|---|---|---|
| NYT Connections (extended) | 65.1% | 30.1% |
| Chess Puzzles | 14% | 4% |
| LMArena Hard Prompts | 1447 | 1448 |
| Mystery Game Puzzles | 8% | 5% |
| DTBench | 78.9% | 82.1% |
| LMCA | 33.7% | 28.3% |
| Epoch Capabilities Index | 146.95 | 142.38 |
| SimpleBench | 45.8% | — |
| Kagi LLM Benchmark | — | 72.5% |
| CritPt | 3.7% | — |
| Surface Evolver Bench | 55% | — |
| ForecastBench | 61.4 | — |

## Math

- MiniMax-M3: 40.0 (#95)
- Qwen3 Max: 38.7 (#131)

| Benchmark | MiniMax-M3 | Qwen3 Max |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 71.1% | 73.3% |
| LMArena Math | 1429 | 1446 |
| FrontierMath (Tiers 1-3) | — | 18.9% |
| ProofBench | 18% | — |
| MATH Level 5 | — | 97.1% |

## Knowledge

- MiniMax-M3: 58.4 (#35)
- Qwen3 Max: 48.1 (#78)

| Benchmark | MiniMax-M3 | Qwen3 Max |
|---|---|---|
| GPQA Diamond | 90.9% | 72.6% |
| LMArena Expert | 1461 | 1455 |
| SimpleQA Verified | — | 48.7% |

## Multimodal

- MiniMax-M3: 40.2 (#51)
- Qwen3 Max: —

| Benchmark | MiniMax-M3 | Qwen3 Max |
|---|---|---|
| LMArena Vision | 1253 | — |
| LMArena Document | 1435 | — |

## Multilingual

- MiniMax-M3: 53.0 (#75)
- Qwen3 Max: 53.7 (#62)

| Benchmark | MiniMax-M3 | Qwen3 Max |
|---|---|---|
| LMArena Non-English | 1420 | 1429 |
| LMArena Chinese | 1463 | 1478 |
| LMArena French | 1447 | 1449 |
| LMArena German | 1426 | 1463 |
| LMArena Japanese | 1381 | 1397 |
| LMArena Korean | 1372 | 1399 |
| LMArena Russian | 1428 | 1428 |
| LMArena Spanish | 1432 | 1462 |

## Instruction Following

- MiniMax-M3: 75.5 (#62)
- Qwen3 Max: 74.8 (#87)

| Benchmark | MiniMax-M3 | Qwen3 Max |
|---|---|---|
| LMArena Instruction Following | 1433 | 1419 |

## Long Context

- MiniMax-M3: 44.2 (#72)
- Qwen3 Max: 41.6 (#134)

| Benchmark | MiniMax-M3 | Qwen3 Max |
|---|---|---|
| LMArena Longer Query | 1445 | 1438 |
| Fiction.LiveBench | — | 66.7% |
| CL-bench | — | 14.5% |

## Writing & Preference

- MiniMax-M3: 62.1 (#83)
- Qwen3 Max: 62.4 (#76)

| Benchmark | MiniMax-M3 | Qwen3 Max |
|---|---|---|
| LMArena Text | 1433 | 1439 |
| LMArena Creative Writing | 1404 | 1402 |
| LMArena Multi-Turn | 1442 | 1446 |
| EQ-Bench 4 | 1150 | — |

## FAQ

### Is MiniMax-M3 better than Qwen3 Max?

MiniMax-M3 and Qwen3 Max score almost the same on the Noometry Index (43.8 vs 43.7), so choose on price, context window or the category you care about most.

### Which is cheaper, MiniMax-M3 or Qwen3 Max?

MiniMax-M3 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; Qwen3 Max lists at $1.20 and $6.

### Is MiniMax-M3 or Qwen3 Max better for coding?

Qwen3 Max scores higher on coding benchmarks: 43.0 versus 41.8 in the Noometry coding category.

### Which has the bigger context window?

MiniMax-M3 does, with 1M tokens against 262K.

### How many benchmarks do MiniMax-M3 and Qwen3 Max share?

27 benchmarks have published results for both models. MiniMax-M3 has 41 scored results on Noometry and Qwen3 Max has 33.