Model comparison
Kimi K3 vs Qwen3 32B
Kimi K3 is the stronger model overall, scoring 59.5 to 39.2 on the Noometry Index. Qwen3 32B costs 4.9× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.
Last verified . 21 shared benchmarks.
Alibaba (Qwen)Summary
- They share 21 benchmarks with published results for both. Kimi K3 scores higher in 9 categories and Qwen3 32B in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Kimi K3 leads 63.0 to 20.2.
- The biggest single-benchmark swing is LMCA: 52.7% for Kimi K3 and 17.3% for Qwen3 32B.
- Qwen3 32B is cheaper at $0.70 / $2.80 per million input/output tokens, against $3 / $15 for Kimi K3.
- Kimi K3 accepts more context: 1.05M tokens versus 131K.
Side by side
| Kimi K3 | Qwen3 32B | |
|---|---|---|
| Provider | Moonshot AI | Alibaba (Qwen) |
| Noometry Index | 59.5 | 39.2 |
| Released | 2026-07-16 | 2025-04 |
| Weights | Open | Open |
| Context window | 1.05M | 131K |
| Max output | 1.05M | 16K |
| Input $ / M tokens | $3 | $0.70 |
| Output $ / M tokens | $15 | $2.80 |
| Results tracked | 53 | 26 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K3 leads
Kimi K3: 61.0 (#10), Qwen3 32B: 37.7 (#190)
| Benchmark | Kimi K3 | Qwen3 32B |
|---|---|---|
| SciCode | 59.5% | 35.4% |
| LMArena Coding | 1508 | 1358 |
| DeepSWE | 68.5% | — |
| FrontierCode | 44.2% | — |
| Aider Polyglot | — | 40% |
| LMArena WebDev | 1654 | — |
| FrontierSWE | 25.9% | — |
| WeirdML | 82.6% | — |
| ALE-Bench | 1,524 | — |
Agentic & Tool Use Kimi K3 leads
Kimi K3: 41.8 (#20), Qwen3 32B: 32.6 (#62)
| Benchmark | Kimi K3 | Qwen3 32B |
|---|---|---|
| APEX-Agents | 50.6% | — |
| Berkeley Function Calling Leaderboard | — | 48.7% |
| τ²-bench Banking | 37.1% | — |
| PostTrainBench | 32% | — |
| GBAEval | 48.3% | — |
| GDP.pdf | 19% | — |
| Vending-Bench 2 | 5,165 | — |
Reasoning Kimi K3 leads
Kimi K3: 63.0 (#17), Qwen3 32B: 20.2 (#241)
| Benchmark | Kimi K3 | Qwen3 32B |
|---|---|---|
| CritPt | 23.4% | 0.3% |
| Chess Puzzles | 39% | 5% |
| LMArena Hard Prompts | 1496 | 1334 |
| DTBench | 91.2% | 67.5% |
| LMCA | 52.7% | 17.3% |
| Epoch Capabilities Index | 157.45 | 138.51 |
| ARC-AGI-2 | 60.4% | — |
| SimpleBench | 60.7% | — |
| Kagi LLM Benchmark | — | 54.9% |
| NYT Connections (extended) | 93.6% | — |
| ARC-AGI-1 | 94.5% | — |
| Mystery Game Puzzles | 26% | — |
| Surface Evolver Bench | 95% | — |
| ForecastBench | 61.1 | — |
Math Kimi K3 leads
Kimi K3: 74.2 (#16), Qwen3 32B: 39.7 (#99)
| Benchmark | Kimi K3 | Qwen3 32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 97.2% | 66.9% |
| LMArena Math | 1491 | 1399 |
| FrontierMath (Tiers 1-3) | 72.2% | — |
| FrontierMath Tier 4 | 39% | — |
| MathArena Final-Answer Competitions | 87.8% | — |
| ProofBench | 87% | — |
Knowledge Kimi K3 leads
Kimi K3: 63.2 (#21), Qwen3 32B: 40.0 (#125)
| Benchmark | Kimi K3 | Qwen3 32B |
|---|---|---|
| GPQA Diamond | 93.1% | 65.7% |
| LMArena Expert | 1521 | 1362 |
| SimpleQA Verified | 50.6% | — |
| Vectara Hallucination Rate | — | 5.9% |
Multimodal Not comparable
Kimi K3: 37.8 (#70), Qwen3 32B: —
| Benchmark | Kimi K3 | Qwen3 32B |
|---|---|---|
| Blueprint-Bench 2 | 29.5% | — |
| Furniture Assembly | 34.2% | — |
Multilingual Kimi K3 leads
Kimi K3: 56.3 (#21), Qwen3 32B: 45.6 (#167)
| Benchmark | Kimi K3 | Qwen3 32B |
|---|---|---|
| LMArena Non-English | 1466 | 1317 |
| LMArena Chinese | 1529 | 1357 |
| LMArena German | 1488 | 1341 |
| LMArena Russian | 1482 | 1311 |
| LMArena French | 1491 | — |
| LMArena Japanese | 1487 | — |
| LMArena Korean | 1458 | — |
| LMArena Spanish | 1472 | — |
Instruction Following Kimi K3 leads
Kimi K3: 77.7 (#14), Qwen3 32B: 68.9 (#179)
| Benchmark | Kimi K3 | Qwen3 32B |
|---|---|---|
| LMArena Instruction Following | 1483 | 1305 |
Long Context Kimi K3 leads
Kimi K3: 45.8 (#29), Qwen3 32B: 43.8 (#87)
| Benchmark | Kimi K3 | Qwen3 32B |
|---|---|---|
| LMArena Longer Query | 1494 | 1327 |
| Fiction.LiveBench | — | 74.2% |
Writing & Preference Kimi K3 leads
Kimi K3: 76.6 (#4), Qwen3 32B: 52.9 (#163)
| Benchmark | Kimi K3 | Qwen3 32B |
|---|---|---|
| LMArena Text | 1476 | 1340 |
| LMArena Creative Writing | 1454 | 1297 |
| LMArena Multi-Turn | 1488 | 1331 |
| EQ-Bench Creative Writing | 2082 | — |
| EQ-Bench 4 | 1339 | — |
Frequently asked questions
Is Kimi K3 better than Qwen3 32B?
Kimi K3 is the stronger model overall, scoring 59.5 to 39.2 on the Noometry Index. Qwen3 32B costs 4.9× 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 32B?
Qwen3 32B is cheaper. It lists at $0.70 per million input tokens and $2.80 per million output tokens; Kimi K3 lists at $3 and $15.
Is Kimi K3 or Qwen3 32B better for coding?
Kimi K3 scores higher on coding benchmarks: 61.0 versus 37.7 in the Noometry coding category.
Which has the bigger context window?
Kimi K3 does, with 1.05M tokens against 131K.
How many benchmarks do Kimi K3 and Qwen3 32B share?
21 benchmarks have published results for both models. Kimi K3 has 53 scored results on Noometry and Qwen3 32B has 26.