# Longcat Flash Chat vs MiniMax-M3

> MiniMax-M3 is the stronger model overall, scoring 43.8 to 42.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/longcat-flash-chat-vs-minimax-m3
- Last updated: 2026-10-11
- Shared benchmarks: 18

## Summary

- They share 18 benchmarks with published results for both. Longcat Flash Chat scores higher in 1 category and MiniMax-M3 in 7 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where MiniMax-M3 leads 58.4 to 40.6.
- The biggest single-benchmark swing is NYT Connections (extended): 17.7% for Longcat Flash Chat and 65.1% for MiniMax-M3.

## Snapshot

| | Longcat Flash Chat | MiniMax-M3 |
|---|---|---|
| Provider | Meituan | MiniMax |
| Noometry Index | 42.1 | 43.8 |
| Rank | 120 | 85 |
| Context | — | 1M |
| Input $/M | — | $0.30 |
| Output $/M | — | $1.20 |
| Weights | Open | Open |

## Coding

- Longcat Flash Chat: 43.5 (#87)
- MiniMax-M3: 41.8 (#118)

| Benchmark | Longcat Flash Chat | MiniMax-M3 |
|---|---|---|
| LMArena Coding | 1471 | 1469 |
| FrontierCode | — | 14.7% |
| LMArena WebDev | — | 1482 |
| SciCode | — | 47.1% |
| ALE-Bench | — | 640.02 |

## Agentic & Tool Use

- Longcat Flash Chat: —
- MiniMax-M3: 22.6 (#130)

| Benchmark | Longcat Flash Chat | MiniMax-M3 |
|---|---|---|
| APEX-Agents | — | 37.7% |
| OSWorld 2.0 | — | 4.6% |
| GBAEval | — | 0.9% |
| Vending-Bench 2 | — | 2,158 |

## Reasoning

- Longcat Flash Chat: 19.0 (#272)
- MiniMax-M3: 30.1 (#87)

| Benchmark | Longcat Flash Chat | MiniMax-M3 |
|---|---|---|
| NYT Connections (extended) | 17.7% | 65.1% |
| LMArena Hard Prompts | 1440 | 1447 |
| SimpleBench | — | 45.8% |
| Kagi LLM Benchmark | 43.9% | — |
| CritPt | — | 3.7% |
| Chess Puzzles | — | 14% |
| Mystery Game Puzzles | — | 8% |
| DTBench | — | 78.9% |
| LMCA | — | 33.7% |
| Surface Evolver Bench | — | 55% |
| Epoch Capabilities Index | — | 146.95 |
| ForecastBench | — | 61.4 |

## Math

- Longcat Flash Chat: 39.4 (#107)
- MiniMax-M3: 40.0 (#95)

| Benchmark | Longcat Flash Chat | MiniMax-M3 |
|---|---|---|
| LMArena Math | 1442 | 1429 |
| OTIS Mock AIME 2024-2025 | — | 71.1% |
| ProofBench | — | 18% |

## Knowledge

- Longcat Flash Chat: 40.6 (#116)
- MiniMax-M3: 58.4 (#35)

| Benchmark | Longcat Flash Chat | MiniMax-M3 |
|---|---|---|
| LMArena Expert | 1454 | 1461 |
| GPQA Diamond | — | 90.9% |

## Multimodal

- Longcat Flash Chat: —
- MiniMax-M3: 40.2 (#51)

| Benchmark | Longcat Flash Chat | MiniMax-M3 |
|---|---|---|
| LMArena Vision | — | 1253 |
| LMArena Document | — | 1435 |

## Multilingual

- Longcat Flash Chat: 51.9 (#101)
- MiniMax-M3: 53.0 (#75)

| Benchmark | Longcat Flash Chat | MiniMax-M3 |
|---|---|---|
| LMArena Non-English | 1404 | 1420 |
| LMArena Chinese | 1465 | 1463 |
| LMArena French | 1456 | 1447 |
| LMArena German | 1408 | 1426 |
| LMArena Japanese | 1373 | 1381 |
| LMArena Korean | 1371 | 1372 |
| LMArena Russian | 1395 | 1428 |
| LMArena Spanish | 1445 | 1432 |

## Instruction Following

- Longcat Flash Chat: 74.4 (#96)
- MiniMax-M3: 75.5 (#62)

| Benchmark | Longcat Flash Chat | MiniMax-M3 |
|---|---|---|
| LMArena Instruction Following | 1411 | 1433 |

## Long Context

- Longcat Flash Chat: 43.5 (#93)
- MiniMax-M3: 44.2 (#72)

| Benchmark | Longcat Flash Chat | MiniMax-M3 |
|---|---|---|
| LMArena Longer Query | 1425 | 1445 |

## Writing & Preference

- Longcat Flash Chat: 61.0 (#91)
- MiniMax-M3: 62.1 (#83)

| Benchmark | Longcat Flash Chat | MiniMax-M3 |
|---|---|---|
| LMArena Text | 1427 | 1433 |
| LMArena Creative Writing | 1388 | 1404 |
| LMArena Multi-Turn | 1418 | 1442 |
| EQ-Bench 4 | — | 1150 |

## FAQ

### Is Longcat Flash Chat better than MiniMax-M3?

MiniMax-M3 is the stronger model overall, scoring 43.8 to 42.1 on the Noometry Index.

### Is Longcat Flash Chat or MiniMax-M3 better for coding?

Longcat Flash Chat scores higher on coding benchmarks: 43.5 versus 41.8 in the Noometry coding category.

### How many benchmarks do Longcat Flash Chat and MiniMax-M3 share?

18 benchmarks have published results for both models. Longcat Flash Chat has 19 scored results on Noometry and MiniMax-M3 has 41.