# Kimi K3 vs Qwen3.5-Flash

> Kimi K3 is the stronger model overall, scoring 59.5 to 42.5 on the Noometry Index. Qwen3.5-Flash costs 34× 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-qwen3-5-flash
- Last updated: 2026-10-10
- Shared benchmarks: 29

## Summary

- They share 29 benchmarks with published results for both. Kimi K3 scores higher in 8 categories and Qwen3.5-Flash in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K3 leads 74.2 to 37.4.
- The biggest single-benchmark swing is FrontierMath (Tiers 1-3): 72.2% for Kimi K3 and 18.2% for Qwen3.5-Flash.
- Qwen3.5-Flash is cheaper at $0.10 / $0.40 per million input/output tokens, against $3 / $15 for Kimi K3.
- Kimi K3 accepts more context: 1.05M tokens versus 1M.
- Kimi K3 has downloadable open weights; the other is API-only.

## Snapshot

| | Kimi K3 | Qwen3.5-Flash |
|---|---|---|
| Provider | Moonshot AI | Alibaba (Qwen) |
| Noometry Index | 59.5 | 42.5 |
| Rank | 15 | 112 |
| Context | 1.05M | 1M |
| Input $/M | $3 | $0.10 |
| Output $/M | $15 | $0.40 |
| Weights | Open | Proprietary |

## Coding

- Kimi K3: 61.0 (#10)
- Qwen3.5-Flash: 34.2 (#242)

| Benchmark | Kimi K3 | Qwen3.5-Flash |
|---|---|---|
| LMArena WebDev | 1654 | 1244 |
| LMArena Coding | 1508 | 1412 |
| ALE-Bench | 1,524 | 221.8 |
| DeepSWE | 68.5% | — |
| FrontierCode | 44.2% | — |
| FrontierSWE | 25.9% | — |
| SciCode | 59.5% | — |
| WeirdML | 82.6% | — |

## Agentic & Tool Use

- Kimi K3: 41.8 (#20)
- Qwen3.5-Flash: —

| Benchmark | Kimi K3 | Qwen3.5-Flash |
|---|---|---|
| Vending-Bench 2 | 5,165 | 462.69 |
| APEX-Agents | 50.6% | — |
| τ²-bench Banking | 37.1% | — |
| PostTrainBench | 32% | — |
| GBAEval | 48.3% | — |
| GDP.pdf | 19% | — |

## Reasoning

- Kimi K3: 63.0 (#17)
- Qwen3.5-Flash: 33.7 (#72)

| Benchmark | Kimi K3 | Qwen3.5-Flash |
|---|---|---|
| Chess Puzzles | 39% | 21% |
| LMArena Hard Prompts | 1496 | 1403 |
| Mystery Game Puzzles | 26% | 20% |
| DTBench | 91.2% | 82.9% |
| LMCA | 52.7% | 29.1% |
| Epoch Capabilities Index | 157.45 | 143.98 |
| ARC-AGI-2 | 60.4% | — |
| SimpleBench | 60.7% | — |
| NYT Connections (extended) | 93.6% | — |
| ARC-AGI-1 | 94.5% | — |
| CritPt | 23.4% | — |
| Surface Evolver Bench | 95% | — |
| ForecastBench | 61.1 | — |

## Math

- Kimi K3: 74.2 (#16)
- Qwen3.5-Flash: 37.4 (#158)

| Benchmark | Kimi K3 | Qwen3.5-Flash |
|---|---|---|
| FrontierMath (Tiers 1-3) | 72.2% | 18.2% |
| OTIS Mock AIME 2024-2025 | 97.2% | 84.4% |
| LMArena Math | 1491 | 1407 |
| FrontierMath Tier 4 | 39% | — |
| MathArena Final-Answer Competitions | 87.8% | — |
| ProofBench | 87% | — |
| FrontierMath (Feb 2025 set) | — | 6.2% |
| FrontierMath Tier 4 (v1) | — | 0% |

## Knowledge

- Kimi K3: 63.2 (#21)
- Qwen3.5-Flash: 43.2 (#93)

| Benchmark | Kimi K3 | Qwen3.5-Flash |
|---|---|---|
| GPQA Diamond | 93.1% | 82.3% |
| SimpleQA Verified | 50.6% | 20.3% |
| LMArena Expert | 1521 | 1407 |
| Vectara Hallucination Rate | — | 10.5% |

## Multimodal

- Kimi K3: 37.8 (#70)
- Qwen3.5-Flash: —

| Benchmark | Kimi K3 | Qwen3.5-Flash |
|---|---|---|
| Blueprint-Bench 2 | 29.5% | — |
| Furniture Assembly | 34.2% | — |

## Multilingual

- Kimi K3: 56.3 (#21)
- Qwen3.5-Flash: 50.5 (#121)

| Benchmark | Kimi K3 | Qwen3.5-Flash |
|---|---|---|
| LMArena Non-English | 1466 | 1385 |
| LMArena Chinese | 1529 | 1446 |
| LMArena French | 1491 | 1412 |
| LMArena German | 1488 | 1390 |
| LMArena Japanese | 1487 | 1368 |
| LMArena Korean | 1458 | 1344 |
| LMArena Russian | 1482 | 1379 |
| LMArena Spanish | 1472 | 1400 |

## Instruction Following

- Kimi K3: 77.7 (#14)
- Qwen3.5-Flash: 72.6 (#139)

| Benchmark | Kimi K3 | Qwen3.5-Flash |
|---|---|---|
| LMArena Instruction Following | 1483 | 1374 |

## Long Context

- Kimi K3: 45.8 (#29)
- Qwen3.5-Flash: 42.4 (#124)

| Benchmark | Kimi K3 | Qwen3.5-Flash |
|---|---|---|
| LMArena Longer Query | 1494 | 1392 |

## Writing & Preference

- Kimi K3: 76.6 (#4)
- Qwen3.5-Flash: 57.9 (#122)

| Benchmark | Kimi K3 | Qwen3.5-Flash |
|---|---|---|
| LMArena Text | 1476 | 1397 |
| LMArena Creative Writing | 1454 | 1343 |
| LMArena Multi-Turn | 1488 | 1393 |
| EQ-Bench Creative Writing | 2082 | — |
| EQ-Bench 4 | 1339 | — |

## FAQ

### Is Kimi K3 better than Qwen3.5-Flash?

Kimi K3 is the stronger model overall, scoring 59.5 to 42.5 on the Noometry Index. Qwen3.5-Flash costs 34× 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 Qwen3.5-Flash?

Qwen3.5-Flash is cheaper. It lists at $0.10 per million input tokens and $0.40 per million output tokens; Kimi K3 lists at $3 and $15.

### Is Kimi K3 or Qwen3.5-Flash better for coding?

Kimi K3 scores higher on coding benchmarks: 61.0 versus 34.2 in the Noometry coding category.

### Which has the bigger context window?

Kimi K3 does, with 1.05M tokens against 1M.

### How many benchmarks do Kimi K3 and Qwen3.5-Flash share?

29 benchmarks have published results for both models. Kimi K3 has 53 scored results on Noometry and Qwen3.5-Flash has 32.