# GPT-4 vs MiniMax-M2

> MiniMax-M2 is the stronger model overall, scoring 37.4 to 29.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gpt-4-vs-minimax-m2
- Last updated: 2026-10-11
- Shared benchmarks: 15

## Summary

- They share 15 benchmarks with published results for both. GPT-4 scores higher in 0 categories and MiniMax-M2 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where MiniMax-M2 leads 37.3 to 10.8.
- MiniMax-M2 is cheaper at $0.30 / $1.20 per million input/output tokens, against $30 / $60 for GPT-4.
- MiniMax-M2 accepts more context: 205K tokens versus 8K.
- MiniMax-M2 has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-4 | MiniMax-M2 |
|---|---|---|
| Provider | OpenAI | MiniMax |
| Noometry Index | 29.1 | 37.4 |
| Rank | 316 | 204 |
| Context | 8K | 205K |
| Input $/M | $30 | $0.30 |
| Output $/M | $60 | $1.20 |
| Weights | Proprietary | Open |

## Coding

- GPT-4: 31.6 (#283)
- MiniMax-M2: 39.3 (#159)

| Benchmark | GPT-4 | MiniMax-M2 |
|---|---|---|
| LMArena Coding | 1254 | 1370 |
| SWE-bench Verified (bash only) | — | 61% |
| LMArena WebDev | — | 1297 |
| WeirdML | 12.4% | — |
| BigCodeBench Instruct | 46% | — |
| BigCodeBench Complete | 57.2% | — |
| HumanEval+ | 79.3% | — |

## Agentic & Tool Use

- GPT-4: —
- MiniMax-M2: 25.1 (#109)

| Benchmark | GPT-4 | MiniMax-M2 |
|---|---|---|
| Terminal-Bench | — | 30% |
| METR Time Horizons | 36.1% | — |
| Vending-Bench 2 | — | 160.6 |

## Reasoning

- GPT-4: 17.8 (#289)
- MiniMax-M2: 19.4 (#258)

| Benchmark | GPT-4 | MiniMax-M2 |
|---|---|---|
| LMArena Hard Prompts | 1241 | 1357 |
| Kagi LLM Benchmark | — | 57.8% |
| NYT Connections (extended) | — | 14.8% |
| Chess Puzzles | 4% | — |
| Mystery Game Puzzles | 12% | — |
| DTBench | 62.7% | — |
| LMCA | 17.1% | — |
| BIG-Bench Hard | 75.1% | — |
| Epoch Capabilities Index | 125.89 | — |
| ForecastBench | 57.8 | — |
| HellaSwag | 95.3% | — |
| WinoGrande | 87.5% | — |

## Math

- GPT-4: 10.8 (#309)
- MiniMax-M2: 37.3 (#160)

| Benchmark | GPT-4 | MiniMax-M2 |
|---|---|---|
| LMArena Math | 1269 | 1352 |
| OTIS Mock AIME 2024-2025 | 1.1% | — |
| MATH Level 5 | 23% | — |
| GSM8K | 92% | — |

## Knowledge

- GPT-4: 18.4 (#282)
- MiniMax-M2: 37.0 (#163)

| Benchmark | GPT-4 | MiniMax-M2 |
|---|---|---|
| LMArena Expert | 1211 | 1337 |
| GPQA Diamond | 35.7% | — |
| MMLU | 86.4% | — |
| TriviaQA | 84.8% | — |

## Multilingual

- GPT-4: 40.6 (#215)
- MiniMax-M2: 45.3 (#171)

| Benchmark | GPT-4 | MiniMax-M2 |
|---|---|---|
| LMArena Non-English | 1246 | 1313 |
| LMArena Chinese | 1242 | 1366 |
| LMArena French | 1283 | 1335 |
| LMArena German | 1251 | 1355 |
| LMArena Russian | 1251 | 1331 |
| LMArena Spanish | 1261 | 1326 |
| LMArena Japanese | 1209 | — |
| LMArena Korean | 1184 | — |

## Instruction Following

- GPT-4: 65.3 (#222)
- MiniMax-M2: 70.2 (#166)

| Benchmark | GPT-4 | MiniMax-M2 |
|---|---|---|
| LMArena Instruction Following | 1241 | 1328 |

## Long Context

- GPT-4: 37.7 (#212)
- MiniMax-M2: 40.5 (#153)

| Benchmark | GPT-4 | MiniMax-M2 |
|---|---|---|
| LMArena Longer Query | 1244 | 1331 |

## Writing & Preference

- GPT-4: 34.9 (#268)
- MiniMax-M2: 53.0 (#162)

| Benchmark | GPT-4 | MiniMax-M2 |
|---|---|---|
| LMArena Text | 1263 | 1340 |
| LMArena Creative Writing | 1244 | 1286 |
| LMArena Multi-Turn | 1257 | 1361 |
| EQ-Bench Creative Writing | 752 | — |

## FAQ

### Is GPT-4 better than MiniMax-M2?

MiniMax-M2 is the stronger model overall, scoring 37.4 to 29.1 on the Noometry Index.

### Which is cheaper, GPT-4 or MiniMax-M2?

MiniMax-M2 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; GPT-4 lists at $30 and $60.

### Is GPT-4 or MiniMax-M2 better for coding?

MiniMax-M2 scores higher on coding benchmarks: 39.3 versus 31.6 in the Noometry coding category.

### Which has the bigger context window?

MiniMax-M2 does, with 205K tokens against 8K.

### How many benchmarks do GPT-4 and MiniMax-M2 share?

15 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and MiniMax-M2 has 21.