# Kimi K3 vs Qwen2.5-Coder-32B

> Kimi K3 is the stronger model overall, scoring 59.5 to 32.9 on the Noometry Index. Qwen2.5-Coder-32B costs 8.1× 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-qwen2-5-coder-32b
- Last updated: 2026-10-10
- Shared benchmarks: 13

## Summary

- They share 13 benchmarks with published results for both. Kimi K3 scores higher in 8 categories and Qwen2.5-Coder-32B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K3 leads 74.0 to 30.8.
- Qwen2.5-Coder-32B is cheaper at $0.66 / $1 per million input/output tokens, against $3 / $15 for Kimi K3.
- Kimi K3 accepts more context: 1.05M tokens versus 33K.

## Snapshot

| | Kimi K3 | Qwen2.5-Coder-32B |
|---|---|---|
| Provider | Moonshot AI | Alibaba (Qwen) |
| Noometry Index | 59.5 | 32.9 |
| Rank | 12 | 239 |
| Context | 1.05M | 33K |
| Input $/M | $3 | $0.66 |
| Output $/M | $15 | $1 |
| Weights | Open | Open |

## Coding

- Kimi K3: 60.6 (#8)
- Qwen2.5-Coder-32B: 22.4 (#328)

| Benchmark | Kimi K3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Coding | 1509 | 1276 |
| DeepSWE | 68.5% | — |
| FrontierCode | 44.2% | — |
| SWE-bench Verified (bash only) | — | 9% |
| Aider Polyglot | — | 16.4% |
| LMArena WebDev | 1659 | — |
| FrontierSWE | 25.9% | — |
| SciCode | 59.5% | — |
| WeirdML | 82.6% | — |
| BigCodeBench Instruct | — | 49% |
| LiveBench Coding | — | 56.9% |
| BigCodeBench Complete | — | 58% |
| ALE-Bench | 1,524 | — |
| HumanEval+ | — | 87.2% |
| MBPP+ | — | 77% |

## Agentic & Tool Use

- Kimi K3: 41.8 (#19)
- Qwen2.5-Coder-32B: —

| Benchmark | Kimi K3 | Qwen2.5-Coder-32B |
|---|---|---|
| APEX-Agents | 50.6% | — |
| τ²-bench Banking | 37.1% | — |
| PostTrainBench | 32% | — |
| GBAEval | 48.3% | — |
| GDP.pdf | 19% | — |
| Vending-Bench 2 | 5,165 | — |

## Reasoning

- Kimi K3: 63.0 (#15)
- Qwen2.5-Coder-32B: 20.4 (#220)

| Benchmark | Kimi K3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Hard Prompts | 1495 | 1251 |
| Epoch Capabilities Index | 157.68 | 119.41 |
| ARC-AGI-2 | 60.4% | — |
| SimpleBench | 60.7% | — |
| NYT Connections (extended) | 93.6% | — |
| ARC-AGI-1 | 94.5% | — |
| CritPt | 23.4% | — |
| Chess Puzzles | 39% | — |
| LiveBench Reasoning | — | 42.1% |
| Mystery Game Puzzles | 26% | — |
| DTBench | 91.2% | — |
| LiveBench Data Analysis | — | 49.9% |
| LMCA | 52.7% | — |
| Surface Evolver Bench | 95% | — |
| ForecastBench | 61.1 | — |
| HellaSwag | — | 83% |
| LiveBench | — | 46.2% |
| WinoGrande | — | 80.8% |

## Math

- Kimi K3: 74.0 (#13)
- Qwen2.5-Coder-32B: 30.8 (#203)

| Benchmark | Kimi K3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Math | 1493 | 1251 |
| FrontierMath (Tiers 1-3) | 72.2% | — |
| FrontierMath Tier 4 | 39% | — |
| MathArena Final-Answer Competitions | 87.8% | — |
| OTIS Mock AIME 2024-2025 | 97.2% | — |
| ProofBench | 87% | — |
| LiveBench Math | — | 46.6% |
| GSM8K | — | 93% |

## Knowledge

- Kimi K3: 63.2 (#19)
- Qwen2.5-Coder-32B: 32.9 (#200)

| Benchmark | Kimi K3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Expert | 1520 | 1221 |
| GPQA Diamond | 93.1% | — |
| SimpleQA Verified | 50.6% | — |
| ARC (AI2) Challenge | — | 70.5% |
| MMLU | — | 79.1% |

## Multimodal

- Kimi K3: 38.4 (#58)
- Qwen2.5-Coder-32B: —

| Benchmark | Kimi K3 | Qwen2.5-Coder-32B |
|---|---|---|
| Blueprint-Bench 2 | 29.5% | — |
| Furniture Assembly | 34.2% | — |

## Multilingual

- Kimi K3: 56.5 (#20)
- Qwen2.5-Coder-32B: 38.0 (#230)

| Benchmark | Kimi K3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Non-English | 1466 | 1205 |
| LMArena Chinese | 1528 | 1222 |
| LMArena Russian | 1481 | 1228 |
| LMArena French | 1488 | — |
| LMArena German | 1474 | — |
| LMArena Japanese | 1492 | — |
| LMArena Korean | 1465 | — |
| LMArena Spanish | 1478 | — |

## Instruction Following

- Kimi K3: 77.8 (#13)
- Qwen2.5-Coder-32B: 61.6 (#240)

| Benchmark | Kimi K3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Instruction Following | 1481 | 1223 |
| LiveBench Instruction Following | — | 58.7% |

## Long Context

- Kimi K3: 45.8 (#27)
- Qwen2.5-Coder-32B: 37.9 (#204)

| Benchmark | Kimi K3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Longer Query | 1494 | 1251 |

## Writing & Preference

- Kimi K3: 76.7 (#4)
- Qwen2.5-Coder-32B: 41.8 (#235)

| Benchmark | Kimi K3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Text | 1476 | 1230 |
| LMArena Creative Writing | 1452 | 1174 |
| LMArena Multi-Turn | 1488 | 1222 |
| EQ-Bench Creative Writing | 2082 | — |
| EQ-Bench 4 | 1339 | — |
| LiveBench Language | — | 23.3% |

## FAQ

### Is Kimi K3 better than Qwen2.5-Coder-32B?

Kimi K3 is the stronger model overall, scoring 59.5 to 32.9 on the Noometry Index. Qwen2.5-Coder-32B costs 8.1× 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 Qwen2.5-Coder-32B?

Qwen2.5-Coder-32B is cheaper. It lists at $0.66 per million input tokens and $1 per million output tokens; Kimi K3 lists at $3 and $15.

### Is Kimi K3 or Qwen2.5-Coder-32B better for coding?

Kimi K3 scores higher on coding benchmarks: 60.6 versus 22.4 in the Noometry coding category.

### Which has the bigger context window?

Kimi K3 does, with 1.05M tokens against 33K.

### How many benchmarks do Kimi K3 and Qwen2.5-Coder-32B share?

13 benchmarks have published results for both models. Kimi K3 has 53 scored results on Noometry and Qwen2.5-Coder-32B has 31.