# DeepSeek-V3 vs MiniMax-M3

> MiniMax-M3 is the stronger model overall, scoring 43.8 to 39.5 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v3-vs-minimax-m3
- Last updated: 2026-10-11
- Shared benchmarks: 26

## Summary

- They share 26 benchmarks with published results for both. DeepSeek-V3 scores higher in 1 category and MiniMax-M3 in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where MiniMax-M3 leads 58.4 to 37.5.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 37.8% for DeepSeek-V3 and 71.1% for MiniMax-M3.
- DeepSeek-V3 is cheaper at $0.24 / $0.90 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M3.
- MiniMax-M3 accepts more context: 1M tokens versus 164K.

## Snapshot

| | DeepSeek-V3 | MiniMax-M3 |
|---|---|---|
| Provider | DeepSeek | MiniMax |
| Noometry Index | 39.5 | 43.8 |
| Rank | 166 | 85 |
| Context | 164K | 1M |
| Input $/M | $0.24 | $0.30 |
| Output $/M | $0.90 | $1.20 |
| Weights | Open | Open |

## Coding

- DeepSeek-V3: 42.3 (#106)
- MiniMax-M3: 41.8 (#118)

| Benchmark | DeepSeek-V3 | MiniMax-M3 |
|---|---|---|
| SciCode | 35.8% | 47.1% |
| LMArena Coding | 1368 | 1469 |
| FrontierCode | — | 14.7% |
| Aider Polyglot | 55.1% | — |
| LMArena WebDev | — | 1482 |
| WeirdML | 36.1% | — |
| BigCodeBench Instruct | 50% | — |
| LiveBench Coding | 70.9% | — |
| BigCodeBench Complete | 62.2% | — |
| ALE-Bench | — | 640.02 |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73% | — |

## Agentic & Tool Use

- DeepSeek-V3: —
- MiniMax-M3: 22.6 (#130)

| Benchmark | DeepSeek-V3 | MiniMax-M3 |
|---|---|---|
| APEX-Agents | — | 37.7% |
| OSWorld 2.0 | — | 4.6% |
| GBAEval | — | 0.9% |
| METR Time Horizons | 49.6% | — |
| Vending-Bench 2 | — | 2,158 |

## Reasoning

- DeepSeek-V3: 20.5 (#236)
- MiniMax-M3: 30.1 (#87)

| Benchmark | DeepSeek-V3 | MiniMax-M3 |
|---|---|---|
| SimpleBench | 27.2% | 45.8% |
| CritPt | 0% | 3.7% |
| LMArena Hard Prompts | 1365 | 1447 |
| DTBench | 64.8% | 78.9% |
| LMCA | 15.5% | 33.7% |
| Epoch Capabilities Index | 135.94 | 146.95 |
| ForecastBench | 59.1 | 61.4 |
| Kagi LLM Benchmark | 52.3% | — |
| NYT Connections (extended) | — | 65.1% |
| Chess Puzzles | — | 14% |
| LiveBench Reasoning | 65.8% | — |
| Mystery Game Puzzles | — | 8% |
| LiveBench Data Analysis | 60.9% | — |
| Surface Evolver Bench | — | 55% |
| BIG-Bench Hard | 87.5% | — |
| HellaSwag | 88.9% | — |
| LiveBench | 66.9% | — |
| PIQA | 84.7% | — |
| WinoGrande | 85.2% | — |

## Math

- DeepSeek-V3: 32.1 (#219)
- MiniMax-M3: 40.0 (#95)

| Benchmark | DeepSeek-V3 | MiniMax-M3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 37.8% | 71.1% |
| LMArena Math | 1373 | 1429 |
| ProofBench | — | 18% |
| Omni-MATH | 40.3% | — |
| LiveBench Math | 73.5% | — |
| MATH Level 5 | 75.5% | — |
| FrontierMath (Feb 2025 set) | 1.7% | — |

## Knowledge

- DeepSeek-V3: 37.5 (#155)
- MiniMax-M3: 58.4 (#35)

| Benchmark | DeepSeek-V3 | MiniMax-M3 |
|---|---|---|
| GPQA Diamond | 67.6% | 90.9% |
| LMArena Expert | 1351 | 1461 |
| MMLU-Pro | 72.3% | — |
| Confabulations | 26.1% | — |
| Vectara Hallucination Rate | 6.1% | — |
| GPQA (HELM) | 53.8% | — |
| ARC (AI2) Challenge | 95.3% | — |
| MMLU | 87.2% | — |
| TriviaQA | 82.9% | — |

## Multimodal

- DeepSeek-V3: —
- MiniMax-M3: 40.2 (#51)

| Benchmark | DeepSeek-V3 | MiniMax-M3 |
|---|---|---|
| LMArena Vision | — | 1253 |
| LMArena Document | — | 1435 |

## Multilingual

- DeepSeek-V3: 48.5 (#143)
- MiniMax-M3: 53.0 (#75)

| Benchmark | DeepSeek-V3 | MiniMax-M3 |
|---|---|---|
| LMArena Non-English | 1358 | 1420 |
| LMArena Chinese | 1391 | 1463 |
| LMArena French | 1385 | 1447 |
| LMArena German | 1374 | 1426 |
| LMArena Japanese | 1333 | 1381 |
| LMArena Korean | 1319 | 1372 |
| LMArena Russian | 1373 | 1428 |
| LMArena Spanish | 1358 | 1432 |

## Instruction Following

- DeepSeek-V3: 72.8 (#130)
- MiniMax-M3: 75.5 (#62)

| Benchmark | DeepSeek-V3 | MiniMax-M3 |
|---|---|---|
| LMArena Instruction Following | 1345 | 1433 |
| LiveBench Instruction Following | 81.5% | — |
| IFEval | 83.2% | — |

## Long Context

- DeepSeek-V3: 34.0 (#253)
- MiniMax-M3: 44.2 (#72)

| Benchmark | DeepSeek-V3 | MiniMax-M3 |
|---|---|---|
| LMArena Longer Query | 1352 | 1445 |
| Fiction.LiveBench | 50% | — |

## Writing & Preference

- DeepSeek-V3: 57.4 (#130)
- MiniMax-M3: 62.1 (#83)

| Benchmark | DeepSeek-V3 | MiniMax-M3 |
|---|---|---|
| LMArena Text | 1375 | 1433 |
| LMArena Creative Writing | 1364 | 1404 |
| LMArena Multi-Turn | 1389 | 1442 |
| Short-Story Creative Writing | 77% | — |
| EQ-Bench Creative Writing | 1472 | — |
| WildBench | 83% | — |
| EQ-Bench 4 | — | 1150 |
| LiveBench Language | 49.1% | — |

## FAQ

### Is DeepSeek-V3 better than MiniMax-M3?

MiniMax-M3 is the stronger model overall, scoring 43.8 to 39.5 on the Noometry Index.

### Which is cheaper, DeepSeek-V3 or MiniMax-M3?

DeepSeek-V3 is cheaper. It lists at $0.24 per million input tokens and $0.90 per million output tokens; MiniMax-M3 lists at $0.30 and $1.20.

### Is DeepSeek-V3 or MiniMax-M3 better for coding?

They score almost the same on coding (42.3 vs 41.8); test both on your own repository before choosing.

### Which has the bigger context window?

MiniMax-M3 does, with 1M tokens against 164K.

### How many benchmarks do DeepSeek-V3 and MiniMax-M3 share?

26 benchmarks have published results for both models. DeepSeek-V3 has 60 scored results on Noometry and MiniMax-M3 has 41.