# Kimi K3 vs MiniMax-M2.5

> Kimi K3 is the stronger model overall, scoring 59.5 to 38.1 on the Noometry Index. MiniMax-M2.5 costs 11× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/kimi-k3-vs-minimax-m2-5
- Last updated: 2026-10-10
- Shared benchmarks: 26

## Summary

- They share 26 benchmarks with published results for both. Kimi K3 scores higher in 9 categories and MiniMax-M2.5 in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K3 leads 74.0 to 25.8.
- The biggest single-benchmark swing is ProofBench: 87% for Kimi K3 and 4% for MiniMax-M2.5.
- MiniMax-M2.5 is cheaper at $0.30 / $1.20 per million input/output tokens, against $3 / $15 for Kimi K3.
- Kimi K3 accepts more context: 1.05M tokens versus 205K.

## Snapshot

| | Kimi K3 | MiniMax-M2.5 |
|---|---|---|
| Provider | Moonshot AI | MiniMax |
| Noometry Index | 59.5 | 38.1 |
| Rank | 12 | 181 |
| Context | 1.05M | 205K |
| Input $/M | $3 | $0.30 |
| Output $/M | $15 | $1.20 |
| Weights | Open | Open |

## Coding

- Kimi K3: 60.6 (#8)
- MiniMax-M2.5: 47.9 (#53)

| Benchmark | Kimi K3 | MiniMax-M2.5 |
|---|---|---|
| LMArena WebDev | 1659 | 1386 |
| LMArena Coding | 1509 | 1380 |
| ALE-Bench | 1,524 | 618.17 |
| DeepSWE | 68.5% | — |
| FrontierCode | 44.2% | — |
| SWE-bench Verified (bash only) | — | 75.8% |
| SWE-bench Multilingual | — | 68.3% |
| FrontierSWE | 25.9% | — |
| SciCode | 59.5% | — |
| WeirdML | 82.6% | — |

## Agentic & Tool Use

- Kimi K3: 41.8 (#19)
- MiniMax-M2.5: 30.2 (#70)

| Benchmark | Kimi K3 | MiniMax-M2.5 |
|---|---|---|
| Vending-Bench 2 | 5,165 | -23.16 |
| Terminal-Bench | — | 42.7% |
| APEX-Agents | 50.6% | — |
| τ²-bench Banking | 37.1% | — |
| PostTrainBench | 32% | — |
| GBAEval | 48.3% | — |
| GDP.pdf | 19% | — |

## Reasoning

- Kimi K3: 63.0 (#15)
- MiniMax-M2.5: 17.6 (#281)

| Benchmark | Kimi K3 | MiniMax-M2.5 |
|---|---|---|
| ARC-AGI-2 | 60.4% | 4.9% |
| NYT Connections (extended) | 93.6% | 16.8% |
| ARC-AGI-1 | 94.5% | 63.7% |
| LMArena Hard Prompts | 1495 | 1372 |
| Epoch Capabilities Index | 157.68 | 146.68 |
| SimpleBench | 60.7% | — |
| Kagi LLM Benchmark | — | 55.2% |
| CritPt | 23.4% | — |
| Chess Puzzles | 39% | — |
| Mystery Game Puzzles | 26% | — |
| DTBench | 91.2% | — |
| LMCA | 52.7% | — |
| Surface Evolver Bench | 95% | — |
| ForecastBench | 61.1 | — |

## Math

- Kimi K3: 74.0 (#13)
- MiniMax-M2.5: 25.8 (#242)

| Benchmark | Kimi K3 | MiniMax-M2.5 |
|---|---|---|
| ProofBench | 87% | 4% |
| LMArena Math | 1493 | 1378 |
| FrontierMath (Tiers 1-3) | 72.2% | — |
| FrontierMath Tier 4 | 39% | — |
| MathArena Final-Answer Competitions | 87.8% | — |
| OTIS Mock AIME 2024-2025 | 97.2% | — |

## Knowledge

- Kimi K3: 63.2 (#19)
- MiniMax-M2.5: 38.6 (#132)

| Benchmark | Kimi K3 | MiniMax-M2.5 |
|---|---|---|
| LMArena Expert | 1520 | 1378 |
| GPQA Diamond | 93.1% | — |
| SimpleQA Verified | 50.6% | — |
| Vectara Hallucination Rate | — | 9.1% |

## Multimodal

- Kimi K3: 38.4 (#58)
- MiniMax-M2.5: —

| Benchmark | Kimi K3 | MiniMax-M2.5 |
|---|---|---|
| Blueprint-Bench 2 | 29.5% | — |
| Furniture Assembly | 34.2% | — |

## Multilingual

- Kimi K3: 56.5 (#20)
- MiniMax-M2.5: 47.3 (#147)

| Benchmark | Kimi K3 | MiniMax-M2.5 |
|---|---|---|
| LMArena Non-English | 1466 | 1338 |
| LMArena Chinese | 1528 | 1392 |
| LMArena French | 1488 | 1363 |
| LMArena German | 1474 | 1361 |
| LMArena Japanese | 1492 | 1169 |
| LMArena Korean | 1465 | 1232 |
| LMArena Russian | 1481 | 1359 |
| LMArena Spanish | 1478 | 1354 |

## Instruction Following

- Kimi K3: 77.8 (#13)
- MiniMax-M2.5: 71.7 (#143)

| Benchmark | Kimi K3 | MiniMax-M2.5 |
|---|---|---|
| LMArena Instruction Following | 1481 | 1353 |

## Long Context

- Kimi K3: 45.8 (#27)
- MiniMax-M2.5: 37.4 (#211)

| Benchmark | Kimi K3 | MiniMax-M2.5 |
|---|---|---|
| LMArena Longer Query | 1494 | 1367 |
| CL-bench | — | 11.4% |
| CL-bench Life | — | 6.3% |

## Writing & Preference

- Kimi K3: 76.7 (#4)
- MiniMax-M2.5: 54.1 (#149)

| Benchmark | Kimi K3 | MiniMax-M2.5 |
|---|---|---|
| LMArena Text | 1476 | 1359 |
| LMArena Creative Writing | 1452 | 1331 |
| EQ-Bench Creative Writing | 2082 | 1361 |
| LMArena Multi-Turn | 1488 | 1364 |
| EQ-Bench 4 | 1339 | — |

## FAQ

### Is Kimi K3 better than MiniMax-M2.5?

Kimi K3 is the stronger model overall, scoring 59.5 to 38.1 on the Noometry Index. MiniMax-M2.5 costs 11× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.

### Which is cheaper, Kimi K3 or MiniMax-M2.5?

MiniMax-M2.5 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; Kimi K3 lists at $3 and $15.

### Is Kimi K3 or MiniMax-M2.5 better for coding?

Kimi K3 scores higher on coding benchmarks: 60.6 versus 47.9 in the Noometry coding category.

### Which has the bigger context window?

Kimi K3 does, with 1.05M tokens against 205K.

### How many benchmarks do Kimi K3 and MiniMax-M2.5 share?

26 benchmarks have published results for both models. Kimi K3 has 53 scored results on Noometry and MiniMax-M2.5 has 33.