# GPT-5.4 vs MiniMax-M2.7

> GPT-5.4 is the stronger model overall, scoring 59.5 to 37.4 on the Noometry Index. MiniMax-M2.7 costs 11× less per token, which makes it the better buy when GPT-5.4's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/gpt-5-4-vs-minimax-m2-7
- Last updated: 2026-10-10
- Shared benchmarks: 29

## Summary

- They share 29 benchmarks with published results for both. GPT-5.4 scores higher in 9 categories and MiniMax-M2.7 in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.4 leads 73.3 to 24.9.
- The biggest single-benchmark swing is NYT Connections (extended): 91.3% for GPT-5.4 and 24.7% for MiniMax-M2.7.
- MiniMax-M2.7 is cheaper at $0.30 / $1.20 per million input/output tokens, against $2.50 / $15 for GPT-5.4.
- GPT-5.4 accepts more context: 1.05M tokens versus 205K.
- MiniMax-M2.7 has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-5.4 | MiniMax-M2.7 |
|---|---|---|
| Provider | OpenAI | MiniMax |
| Noometry Index | 59.5 | 37.4 |
| Rank | 14 | 188 |
| Context | 1.05M | 205K |
| Input $/M | $2.50 | $0.30 |
| Output $/M | $15 | $1.20 |
| Weights | Proprietary | Open |

## Coding

- GPT-5.4: 52.8 (#28)
- MiniMax-M2.7: 41.4 (#115)

| Benchmark | GPT-5.4 | MiniMax-M2.7 |
|---|---|---|
| LMArena WebDev | 1465 | 1397 |
| SciCode | 56.6% | 47% |
| WeirdML | 77.7% | 37% |
| LMArena Coding | 1497 | 1454 |
| ALE-Bench | 1,607 | 599.25 |
| SWE-bench Verified | 76.9% | — |
| DeepSWE | 51.8% | — |
| GSO | 31.4% | — |
| MirrorCode | 15.6% | — |
| AlgoTune | 1.85 | — |

## Agentic & Tool Use

- GPT-5.4: 46.5 (#12)
- MiniMax-M2.7: 24.9 (#103)

| Benchmark | GPT-5.4 | MiniMax-M2.7 |
|---|---|---|
| Terminal-Bench | 81.8% | 45.1% |
| GBAEval | 45.1% | 0% |
| APEX-Agents | 52.4% | — |
| τ²-bench Banking | 39.4% | — |
| DeepResearch Bench | 35.1% | — |
| PostTrainBench | 19% | — |
| ExploitBench | — | 13.3% |
| LMArena Search | 1197 | — |
| METR Time Horizons | 74.3% | — |
| Vending-Bench 2 | 6,144 | — |

## Reasoning

- GPT-5.4: 61.7 (#17)
- MiniMax-M2.7: 19.0 (#248)

| Benchmark | GPT-5.4 | MiniMax-M2.7 |
|---|---|---|
| NYT Connections (extended) | 91.3% | 24.7% |
| CritPt | 23.4% | 0.6% |
| Thematic Generalization | 80% | 39.3% |
| LMArena Hard Prompts | 1485 | 1422 |
| Epoch Capabilities Index | 156.92 | 145.84 |
| ARC-AGI-2 | 74% | — |
| Kagi LLM Benchmark | 63.8% | — |
| ARC-AGI-1 | 93.7% | — |
| Chess Puzzles | 44% | — |
| EnigmaEval | 16% | — |
| EBR-Bench | 25.4% | — |
| Mystery Game Puzzles | 37% | — |
| DTBench | 94.4% | — |
| LMCA | 52% | — |
| ForecastBench | 59.5 | — |

## Math

- GPT-5.4: 73.3 (#14)
- MiniMax-M2.7: 24.9 (#254)

| Benchmark | GPT-5.4 | MiniMax-M2.7 |
|---|---|---|
| ProofBench | 56% | 3% |
| LMArena Math | 1488 | 1420 |
| FrontierMath (Tiers 1-3) | 78.6% | — |
| FrontierMath Tier 4 | 49% | — |
| MathArena Final-Answer Competitions | 83.1% | — |
| OTIS Mock AIME 2024-2025 | 97.8% | — |
| FrontierMath (Feb 2025 set) | 47.6% | — |
| FrontierMath Tier 4 (v1) | 27.1% | — |

## Knowledge

- GPT-5.4: 65.3 (#13)
- MiniMax-M2.7: 37.1 (#151)

| Benchmark | GPT-5.4 | MiniMax-M2.7 |
|---|---|---|
| Vectara Hallucination Rate | 7% | 12.9% |
| LMArena Expert | 1507 | 1443 |
| GPQA Diamond | 93.3% | — |
| Humanity's Last Exam | 36.2% | — |
| SimpleQA Verified | 45.1% | — |

## Multimodal

- GPT-5.4: 43.8 (#19)
- MiniMax-M2.7: —

| Benchmark | GPT-5.4 | MiniMax-M2.7 |
|---|---|---|
| LMArena Vision | 1302 | — |
| Blueprint-Bench 2 | 27.1% | — |
| Furniture Assembly | 37.5% | — |
| LMArena Document | 1471 | — |

## Multilingual

- GPT-5.4: 56.4 (#22)
- MiniMax-M2.7: 50.5 (#120)

| Benchmark | GPT-5.4 | MiniMax-M2.7 |
|---|---|---|
| LMArena Non-English | 1465 | 1382 |
| LMArena Chinese | 1520 | 1439 |
| LMArena French | 1493 | 1422 |
| LMArena German | 1472 | 1398 |
| LMArena Japanese | 1484 | 1261 |
| LMArena Korean | 1449 | 1313 |
| LMArena Russian | 1480 | 1383 |
| LMArena Spanish | 1453 | 1403 |

## Instruction Following

- GPT-5.4: 77.3 (#25)
- MiniMax-M2.7: 74.2 (#101)

| Benchmark | GPT-5.4 | MiniMax-M2.7 |
|---|---|---|
| LMArena Instruction Following | 1469 | 1404 |

## Long Context

- GPT-5.4: 50.3 (#8)
- MiniMax-M2.7: 43.3 (#94)

| Benchmark | GPT-5.4 | MiniMax-M2.7 |
|---|---|---|
| LMArena Longer Query | 1474 | 1419 |
| CL-bench | 27.9% | — |
| CL-bench Life | 21.7% | — |

## Writing & Preference

- GPT-5.4: 72.1 (#18)
- MiniMax-M2.7: 59.1 (#107)

| Benchmark | GPT-5.4 | MiniMax-M2.7 |
|---|---|---|
| LMArena Text | 1469 | 1404 |
| LMArena Creative Writing | 1439 | 1354 |
| LMArena Multi-Turn | 1482 | 1413 |
| EQ-Bench Creative Writing | 1840 | — |
| EQ-Bench 4 | 1272 | — |

## FAQ

### Is GPT-5.4 better than MiniMax-M2.7?

GPT-5.4 is the stronger model overall, scoring 59.5 to 37.4 on the Noometry Index. MiniMax-M2.7 costs 11× less per token, which makes it the better buy when GPT-5.4's lead doesn't matter for your workload.

### Which is cheaper, GPT-5.4 or MiniMax-M2.7?

MiniMax-M2.7 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; GPT-5.4 lists at $2.50 and $15.

### Is GPT-5.4 or MiniMax-M2.7 better for coding?

GPT-5.4 scores higher on coding benchmarks: 52.8 versus 41.4 in the Noometry coding category.

### Which has the bigger context window?

GPT-5.4 does, with 1.05M tokens against 205K.

### How many benchmarks do GPT-5.4 and MiniMax-M2.7 share?

29 benchmarks have published results for both models. GPT-5.4 has 68 scored results on Noometry and MiniMax-M2.7 has 30.