Model comparison
Kimi K3 vs Qwen3 235B-A22B
Kimi K3 is the stronger model overall, scoring 59.5 to 43.2 on the Noometry Index. Qwen3 235B-A22B 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 . 34 shared benchmarks.
Alibaba (Qwen)Summary
- They share 34 benchmarks with published results for both. Kimi K3 scores higher in 8 categories and Qwen3 235B-A22B in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Kimi K3 leads 63.0 to 15.8.
- The biggest single-benchmark swing is ARC-AGI-1: 94.5% for Kimi K3 and 11% for Qwen3 235B-A22B.
- Qwen3 235B-A22B 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 235B-A22B | |
|---|---|---|
| Provider | Moonshot AI | Alibaba (Qwen) |
| Noometry Index | 59.5 | 43.2 |
| 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 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K3 leads
Kimi K3: 60.6 (#8), Qwen3 235B-A22B: 43.8 (#72)
| Benchmark | Kimi K3 | Qwen3 235B-A22B |
|---|---|---|
| SciCode | 59.5% | 42.4% |
| WeirdML | 82.6% | 41% |
| LMArena Coding | 1509 | 1445 |
| DeepSWE | 68.5% | — |
| FrontierCode | 44.2% | — |
| Aider Polyglot | — | 59.6% |
| LMArena WebDev | 1659 | — |
| FrontierSWE | 25.9% | — |
| ALE-Bench | 1,524 | — |
Agentic & Tool Use Kimi K3 leads
Kimi K3: 41.8 (#19), Qwen3 235B-A22B: 33.7 (#48)
| Benchmark | Kimi K3 | Qwen3 235B-A22B |
|---|---|---|
| Vending-Bench 2 | 5,165 | -11.34 |
| APEX-Agents | 50.6% | — |
| Berkeley Function Calling Leaderboard | — | 52.1% |
| τ²-bench Banking | 37.1% | — |
| PostTrainBench | 32% | — |
| GBAEval | 48.3% | — |
| GDP.pdf | 19% | — |
Reasoning Kimi K3 leads
Kimi K3: 63.0 (#15), Qwen3 235B-A22B: 15.8 (#302)
| Benchmark | Kimi K3 | Qwen3 235B-A22B |
|---|---|---|
| ARC-AGI-2 | 60.4% | 1.3% |
| SimpleBench | 60.7% | 31% |
| ARC-AGI-1 | 94.5% | 11% |
| CritPt | 23.4% | 0% |
| Chess Puzzles | 39% | 12% |
| LMArena Hard Prompts | 1495 | 1433 |
| Mystery Game Puzzles | 26% | 9% |
| DTBench | 91.2% | 80.3% |
| LMCA | 52.7% | 29.3% |
| Epoch Capabilities Index | 157.68 | 143.87 |
| ForecastBench | 61.1 | 59.7 |
| Kagi LLM Benchmark | — | 69.4% |
| NYT Connections (extended) | 93.6% | — |
| Surface Evolver Bench | 95% | — |
Math Kimi K3 leads
Kimi K3: 74.0 (#13), Qwen3 235B-A22B: 48.5 (#51)
| Benchmark | Kimi K3 | Qwen3 235B-A22B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 97.2% | 86.7% |
| LMArena Math | 1493 | 1431 |
| FrontierMath (Tiers 1-3) | 72.2% | — |
| FrontierMath Tier 4 | 39% | — |
| MathArena Final-Answer Competitions | 87.8% | — |
| ProofBench | 87% | — |
| Omni-MATH | — | 71.8% |
| MATH Level 5 | — | 68.9% |
| FrontierMath (Feb 2025 set) | — | 8.5% |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge Kimi K3 leads
Kimi K3: 63.2 (#19), Qwen3 235B-A22B: 49.5 (#67)
| Benchmark | Kimi K3 | Qwen3 235B-A22B |
|---|---|---|
| GPQA Diamond | 93.1% | 80.1% |
| SimpleQA Verified | 50.6% | 40.4% |
| LMArena Expert | 1520 | 1460 |
| MMLU-Pro | — | 84.4% |
| Confabulations | — | 15.6% |
| Vectara Hallucination Rate | — | 9.3% |
| GPQA (HELM) | — | 72.7% |
Multimodal Not comparable
Kimi K3: 38.4 (#58), Qwen3 235B-A22B: —
| Benchmark | Kimi K3 | Qwen3 235B-A22B |
|---|---|---|
| Blueprint-Bench 2 | 29.5% | — |
| Furniture Assembly | 34.2% | — |
Multilingual Kimi K3 leads
Kimi K3: 56.5 (#20), Qwen3 235B-A22B: 52.5 (#84)
| Benchmark | Kimi K3 | Qwen3 235B-A22B |
|---|---|---|
| LMArena Non-English | 1466 | 1410 |
| LMArena Chinese | 1528 | 1481 |
| LMArena French | 1488 | 1445 |
| LMArena German | 1474 | 1432 |
| LMArena Japanese | 1492 | 1400 |
| LMArena Korean | 1465 | 1391 |
| LMArena Russian | 1481 | 1411 |
| LMArena Spanish | 1478 | 1429 |
Instruction Following Kimi K3 leads
Kimi K3: 77.8 (#13), Qwen3 235B-A22B: 72.8 (#132)
| Benchmark | Kimi K3 | Qwen3 235B-A22B |
|---|---|---|
| LMArena Instruction Following | 1481 | 1408 |
| IFEval | — | 83.5% |
Long Context Too close to call
Kimi K3: 45.8 (#27), Qwen3 235B-A22B: 46.1 (#25)
| Benchmark | Kimi K3 | Qwen3 235B-A22B |
|---|---|---|
| LMArena Longer Query | 1494 | 1426 |
| Fiction.LiveBench | — | 75% |
Writing & Preference Kimi K3 leads
Kimi K3: 76.7 (#4), Qwen3 235B-A22B: 59.8 (#103)
| Benchmark | Kimi K3 | Qwen3 235B-A22B |
|---|---|---|
| LMArena Text | 1476 | 1419 |
| LMArena Creative Writing | 1452 | 1384 |
| EQ-Bench Creative Writing | 2082 | 1366 |
| LMArena Multi-Turn | 1488 | 1432 |
| Short-Story Creative Writing | — | 83% |
| WildBench | — | 86.6% |
| EQ-Bench 4 | 1339 | — |
Frequently asked questions
Is Kimi K3 better than Qwen3 235B-A22B?
Kimi K3 is the stronger model overall, scoring 59.5 to 43.2 on the Noometry Index. Qwen3 235B-A22B 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 235B-A22B?
Qwen3 235B-A22B 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 235B-A22B better for coding?
Kimi K3 scores higher on coding benchmarks: 60.6 versus 43.8 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 235B-A22B share?
34 benchmarks have published results for both models. Kimi K3 has 53 scored results on Noometry and Qwen3 235B-A22B has 49.