# Llama 4 Maverick vs MiniMax-M3

> MiniMax-M3 is the stronger model overall, scoring 43.5 to 30.6 on the Noometry Index. Llama 4 Maverick costs 1.7× less per token, which makes it the better buy when MiniMax-M3's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/llama-4-maverick-vs-minimax-m3
- Last updated: 2026-10-10
- Shared benchmarks: 29

## Summary

- They share 29 benchmarks with published results for both. Llama 4 Maverick scores higher in 1 category and MiniMax-M3 in 9 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where MiniMax-M3 leads 58.4 to 33.3.
- The biggest single-benchmark swing is NYT Connections (extended): 8% for Llama 4 Maverick and 65.1% for MiniMax-M3.
- Llama 4 Maverick is cheaper at $0.19 / $0.65 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M3.
- MiniMax-M3 accepts more context: 1M tokens versus 128K.

## Snapshot

| | Llama 4 Maverick | MiniMax-M3 |
|---|---|---|
| Provider | Meta | MiniMax |
| Noometry Index | 30.6 | 43.5 |
| Rank | 276 | 81 |
| Context | 128K | 1M |
| Input $/M | $0.19 | $0.30 |
| Output $/M | $0.65 | $1.20 |
| Weights | Open | Open |

## Coding

- Llama 4 Maverick: 26.3 (#318)
- MiniMax-M3: 41.5 (#111)

| Benchmark | Llama 4 Maverick | MiniMax-M3 |
|---|---|---|
| SciCode | 33.1% | 47.1% |
| LMArena Coding | 1302 | 1470 |
| ALE-Bench | 172.97 | 640.02 |
| FrontierCode | — | 14.7% |
| SWE-bench Verified (bash only) | 21% | — |
| Aider Polyglot | 15.6% | — |
| LMArena WebDev | — | 1482 |
| WeirdML | 24.5% | — |
| BigCodeBench Instruct | 49.7% | — |
| BigCodeBench Complete | 61.4% | — |

## Agentic & Tool Use

- Llama 4 Maverick: 27.9 (#86)
- MiniMax-M3: 22.5 (#121)

| Benchmark | Llama 4 Maverick | MiniMax-M3 |
|---|---|---|
| APEX-Agents | — | 37.7% |
| Berkeley Function Calling Leaderboard | 37.3% | — |
| OSWorld 2.0 | — | 4.6% |
| GBAEval | — | 0.9% |
| Vending-Bench 2 | — | 2,158 |

## Reasoning

- Llama 4 Maverick: 10.1 (#337)
- MiniMax-M3: 29.3 (#83)

| Benchmark | Llama 4 Maverick | MiniMax-M3 |
|---|---|---|
| SimpleBench | 27.7% | 45.8% |
| NYT Connections (extended) | 8% | 65.1% |
| CritPt | 0% | 3.7% |
| LMArena Hard Prompts | 1281 | 1447 |
| DTBench | 61.9% | 78.9% |
| LMCA | 15.9% | 33.7% |
| Epoch Capabilities Index | 132.2 | 147 |
| ForecastBench | 57.5 | 61.4 |
| ARC-AGI-2 | 0% | — |
| Kagi LLM Benchmark | 55.9% | — |
| ARC-AGI-1 | 4.4% | — |
| Chess Puzzles | — | 14% |
| EnigmaEval | 0.6% | — |
| Mystery Game Puzzles | — | 8% |
| Surface Evolver Bench | — | 55% |

## Math

- Llama 4 Maverick: 24.6 (#256)
- MiniMax-M3: 39.1 (#83)

| Benchmark | Llama 4 Maverick | MiniMax-M3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 20.6% | 71.1% |
| LMArena Math | 1299 | 1432 |
| ProofBench | — | 18% |
| Omni-MATH | 42.2% | — |
| MATH Level 5 | 73% | — |
| FrontierMath (Feb 2025 set) | 0.7% | — |

## Knowledge

- Llama 4 Maverick: 33.3 (#196)
- MiniMax-M3: 58.4 (#32)

| Benchmark | Llama 4 Maverick | MiniMax-M3 |
|---|---|---|
| GPQA Diamond | 67% | 90.9% |
| LMArena Expert | 1259 | 1462 |
| Humanity's Last Exam | 5.7% | — |
| MMLU-Pro | 81% | — |
| Confabulations | 22.6% | — |
| Vectara Hallucination Rate | 8.2% | — |
| GPQA (HELM) | 65% | — |

## Multimodal

- Llama 4 Maverick: 30.9 (#100)
- MiniMax-M3: 39.5 (#45)

| Benchmark | Llama 4 Maverick | MiniMax-M3 |
|---|---|---|
| LMArena Vision | 1141 | 1254 |
| GeoBench | 52% | — |
| LMArena Document | — | 1435 |
| SpatialViz-Bench | 31.8% | — |

## Multilingual

- Llama 4 Maverick: 42.4 (#190)
- MiniMax-M3: 53.2 (#70)

| Benchmark | Llama 4 Maverick | MiniMax-M3 |
|---|---|---|
| LMArena Non-English | 1269 | 1420 |
| LMArena Chinese | 1277 | 1465 |
| LMArena French | 1258 | 1445 |
| LMArena German | 1291 | 1424 |
| LMArena Japanese | 1207 | 1381 |
| LMArena Korean | 1203 | 1372 |
| LMArena Russian | 1286 | 1430 |
| LMArena Spanish | 1293 | 1431 |

## Instruction Following

- Llama 4 Maverick: 71.9 (#140)
- MiniMax-M3: 75.7 (#56)

| Benchmark | Llama 4 Maverick | MiniMax-M3 |
|---|---|---|
| LMArena Instruction Following | 1267 | 1433 |
| IFEval | 90.8% | — |

## Long Context

- Llama 4 Maverick: 31.4 (#274)
- MiniMax-M3: 44.2 (#68)

| Benchmark | Llama 4 Maverick | MiniMax-M3 |
|---|---|---|
| LMArena Longer Query | 1280 | 1445 |
| Fiction.LiveBench | 46.2% | — |

## Writing & Preference

- Llama 4 Maverick: 39.0 (#247)
- MiniMax-M3: 62.3 (#79)

| Benchmark | Llama 4 Maverick | MiniMax-M3 |
|---|---|---|
| LMArena Text | 1287 | 1433 |
| LMArena Creative Writing | 1266 | 1405 |
| LMArena Multi-Turn | 1289 | 1441 |
| Short-Story Creative Writing | 62% | — |
| EQ-Bench Creative Writing | 860 | — |
| WildBench | 80% | — |
| EQ-Bench 4 | — | 1150 |

## FAQ

### Is Llama 4 Maverick better than MiniMax-M3?

MiniMax-M3 is the stronger model overall, scoring 43.5 to 30.6 on the Noometry Index. Llama 4 Maverick costs 1.7× less per token, which makes it the better buy when MiniMax-M3's lead doesn't matter for your workload.

### Which is cheaper, Llama 4 Maverick or MiniMax-M3?

Llama 4 Maverick is cheaper. It lists at $0.19 per million input tokens and $0.65 per million output tokens; MiniMax-M3 lists at $0.30 and $1.20.

### Is Llama 4 Maverick or MiniMax-M3 better for coding?

MiniMax-M3 scores higher on coding benchmarks: 41.5 versus 26.3 in the Noometry coding category.

### Which has the bigger context window?

MiniMax-M3 does, with 1M tokens against 128K.

### How many benchmarks do Llama 4 Maverick and MiniMax-M3 share?

29 benchmarks have published results for both models. Llama 4 Maverick has 54 scored results on Noometry and MiniMax-M3 has 41.