# Claude Sonnet 4.6 vs MiniMax-M3

> Claude Sonnet 4.6 is the stronger model overall, scoring 50.3 to 43.8 on the Noometry Index. MiniMax-M3 costs 11× less per token, which makes it the better buy when Claude Sonnet 4.6's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/claude-sonnet-4-6-vs-minimax-m3
- Last updated: 2026-10-12
- Shared benchmarks: 39

## Summary

- They share 39 benchmarks with published results for both. Claude Sonnet 4.6 scores higher in 8 categories and MiniMax-M3 in 2 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where Claude Sonnet 4.6 leads 39.1 to 22.6.
- The biggest single-benchmark swing is GBAEval: 48.8% for Claude Sonnet 4.6 and 0.9% for MiniMax-M3.
- MiniMax-M3 is cheaper at $0.30 / $1.20 per million input/output tokens, against $3 / $15 for Claude Sonnet 4.6.
- MiniMax-M3 has downloadable open weights; the other is API-only.

## Snapshot

| | Claude Sonnet 4.6 | MiniMax-M3 |
|---|---|---|
| Provider | Anthropic | MiniMax |
| Noometry Index | 50.3 | 43.8 |
| Rank | 50 | 85 |
| Context | 1M | 1M |
| Input $/M | $3 | $0.30 |
| Output $/M | $15 | $1.20 |
| Weights | Proprietary | Open |

## Coding

- Claude Sonnet 4.6: 46.3 (#67)
- MiniMax-M3: 41.8 (#118)

| Benchmark | Claude Sonnet 4.6 | MiniMax-M3 |
|---|---|---|
| FrontierCode | 24.3% | 14.7% |
| LMArena WebDev | 1522 | 1482 |
| SciCode | 46.8% | 47.1% |
| LMArena Coding | 1504 | 1469 |
| ALE-Bench | 1,327 | 640.02 |
| SWE-bench Verified | 75.2% | — |
| DeepSWE | 29.9% | — |
| WeirdML | 66.1% | — |

## Agentic & Tool Use

- Claude Sonnet 4.6: 39.1 (#28)
- MiniMax-M3: 22.6 (#130)

| Benchmark | Claude Sonnet 4.6 | MiniMax-M3 |
|---|---|---|
| APEX-Agents | 43% | 37.7% |
| OSWorld 2.0 | 9.3% | 4.6% |
| GBAEval | 48.8% | 0.9% |
| Vending-Bench 2 | 7,204 | 2,158 |
| Terminal-Bench | 53.4% | — |
| DeepResearch Bench | 54.9% | — |
| OSWorld | 72.1% | — |
| ExploitBench | 23.6% | — |
| GDP.pdf | 18% | — |
| LMArena Search | 1221 | — |

## Reasoning

- Claude Sonnet 4.6: 46.1 (#45)
- MiniMax-M3: 30.1 (#87)

| Benchmark | Claude Sonnet 4.6 | MiniMax-M3 |
|---|---|---|
| NYT Connections (extended) | 80.9% | 65.1% |
| CritPt | 3.1% | 3.7% |
| Chess Puzzles | 13% | 14% |
| LMArena Hard Prompts | 1484 | 1447 |
| Mystery Game Puzzles | 16% | 8% |
| DTBench | 89.9% | 78.9% |
| LMCA | 46.5% | 33.7% |
| Epoch Capabilities Index | 152.24 | 146.95 |
| ForecastBench | 62 | 61.4 |
| ARC-AGI-2 | 60.4% | — |
| SimpleBench | — | 45.8% |
| ARC-AGI-1 | 86.5% | — |
| Thematic Generalization | 76.3% | — |
| Surface Evolver Bench | — | 55% |

## Math

- Claude Sonnet 4.6: 52.9 (#49)
- MiniMax-M3: 40.0 (#95)

| Benchmark | Claude Sonnet 4.6 | MiniMax-M3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 85.8% | 71.1% |
| ProofBench | 45% | 18% |
| LMArena Math | 1462 | 1429 |
| FrontierMath (Feb 2025 set) | 32.4% | — |
| FrontierMath Tier 4 (v1) | 8.3% | — |

## Knowledge

- Claude Sonnet 4.6: 51.7 (#65)
- MiniMax-M3: 58.4 (#35)

| Benchmark | Claude Sonnet 4.6 | MiniMax-M3 |
|---|---|---|
| GPQA Diamond | 87.4% | 90.9% |
| LMArena Expert | 1500 | 1461 |
| SimpleQA Verified | 35.5% | — |
| Vectara Hallucination Rate | 10.6% | — |

## Multimodal

- Claude Sonnet 4.6: 38.0 (#68)
- MiniMax-M3: 40.2 (#51)

| Benchmark | Claude Sonnet 4.6 | MiniMax-M3 |
|---|---|---|
| LMArena Vision | 1283 | 1253 |
| LMArena Document | 1482 | 1435 |
| Blueprint-Bench 2 | 6.7% | — |

## Multilingual

- Claude Sonnet 4.6: 54.4 (#41)
- MiniMax-M3: 53.0 (#75)

| Benchmark | Claude Sonnet 4.6 | MiniMax-M3 |
|---|---|---|
| LMArena Non-English | 1440 | 1420 |
| LMArena Chinese | 1491 | 1463 |
| LMArena French | 1465 | 1447 |
| LMArena German | 1428 | 1426 |
| LMArena Japanese | 1420 | 1381 |
| LMArena Korean | 1411 | 1372 |
| LMArena Russian | 1440 | 1428 |
| LMArena Spanish | 1464 | 1432 |

## Instruction Following

- Claude Sonnet 4.6: 77.4 (#25)
- MiniMax-M3: 75.5 (#62)

| Benchmark | Claude Sonnet 4.6 | MiniMax-M3 |
|---|---|---|
| LMArena Instruction Following | 1475 | 1433 |

## Long Context

- Claude Sonnet 4.6: 45.3 (#44)
- MiniMax-M3: 44.2 (#72)

| Benchmark | Claude Sonnet 4.6 | MiniMax-M3 |
|---|---|---|
| LMArena Longer Query | 1479 | 1445 |

## Writing & Preference

- Claude Sonnet 4.6: 70.2 (#22)
- MiniMax-M3: 62.1 (#83)

| Benchmark | Claude Sonnet 4.6 | MiniMax-M3 |
|---|---|---|
| LMArena Text | 1458 | 1433 |
| LMArena Creative Writing | 1435 | 1404 |
| EQ-Bench 4 | 1207 | 1150 |
| LMArena Multi-Turn | 1464 | 1442 |
| EQ-Bench Creative Writing | 1810 | — |

## FAQ

### Is Claude Sonnet 4.6 better than MiniMax-M3?

Claude Sonnet 4.6 is the stronger model overall, scoring 50.3 to 43.8 on the Noometry Index. MiniMax-M3 costs 11× less per token, which makes it the better buy when Claude Sonnet 4.6's lead doesn't matter for your workload.

### Which is cheaper, Claude Sonnet 4.6 or MiniMax-M3?

MiniMax-M3 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; Claude Sonnet 4.6 lists at $3 and $15.

### Is Claude Sonnet 4.6 or MiniMax-M3 better for coding?

Claude Sonnet 4.6 scores higher on coding benchmarks: 46.3 versus 41.8 in the Noometry coding category.

### Which has the bigger context window?

Both accept 1M tokens.

### How many benchmarks do Claude Sonnet 4.6 and MiniMax-M3 share?

39 benchmarks have published results for both models. Claude Sonnet 4.6 has 57 scored results on Noometry and MiniMax-M3 has 41.