Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.
Kimi K2.5
71
Winner · 3/8 categoriesQwen3.5-35B-A3B
67
4/8 categoriesKimi K2.5· Qwen3.5-35B-A3B
Pick Kimi K2.5 if you want the stronger benchmark profile. Qwen3.5-35B-A3B only becomes the better choice if knowledge is the priority or you want the cheaper token bill.
Kimi K2.5 is clearly ahead on the aggregate, 71 to 67. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
Kimi K2.5's sharpest advantage is in reasoning, where it averages 69.3 against 59. The single biggest benchmark swing on the page is SuperGPQA, 74% to 63.4%. Qwen3.5-35B-A3B does hit back in knowledge, so the answer changes if that is the part of the workload you care about most.
Kimi K2.5 is also the more expensive model on tokens at $0.50 input / $2.80 output per 1M tokens, versus $0.00 input / $0.00 output per 1M tokens for Qwen3.5-35B-A3B. That is roughly Infinityx on output cost alone. Qwen3.5-35B-A3B is the reasoning model in the pair, while Kimi K2.5 is not. That usually helps on harder chain-of-thought-heavy tests, but it can also mean more latency and more token spend in real use. Qwen3.5-35B-A3B gives you the larger context window at 262K, compared with 128K for Kimi K2.5.
BenchLM keeps the benchmark table and the operator tradeoffs on the same page so a better score does not hide a materially slower, pricier, or smaller-context model.
Runtime metrics show N/A when BenchLM does not have a sourced snapshot for that exact model. The scoring rules and freshness policy are documented on the methodology page.
| Benchmark | Kimi K2.5 | Qwen3.5-35B-A3B |
|---|---|---|
| AgenticKimi K2.5 wins | ||
| Terminal-Bench 2.0 | 50.8% | 40.5% |
| BrowseComp | 60.6% | 61% |
| OSWorld-Verified | 63.3% | 54.5% |
| tau2-bench | — | 81.2% |
| CodingQwen3.5-35B-A3B wins | ||
| HumanEval | 99% | — |
| SWE-bench Verified | 76.8% | 69.2% |
| LiveCodeBench | 85% | 74.6% |
| SWE-bench Pro | 40% | — |
| SWE-Rebench | 58.5% | — |
| React Native Evals | 74.9% | — |
| Multimodal & GroundedQwen3.5-35B-A3B wins | ||
| MMMU-Pro | 78.5% | 75.1% |
| OfficeQA Pro | 69% | — |
| ReasoningKimi K2.5 wins | ||
| MuSR | 72% | — |
| BBH | 81% | — |
| LongBench v2 | 67% | 59% |
| MRCRv2 | 70% | — |
| KnowledgeQwen3.5-35B-A3B wins | ||
| MMLU | 77% | — |
| GPQA | 87.6% | 84.2% |
| SuperGPQA | 74% | 63.4% |
| MMLU-Pro | 87.1% | 85.3% |
| HLE | 11% | — |
| FrontierScience | 67% | — |
| SimpleQA | 74% | — |
| Instruction FollowingKimi K2.5 wins | ||
| IFEval | 94% | 91.9% |
| MultilingualQwen3.5-35B-A3B wins | ||
| MGSM | 83% | — |
| MMLU-ProX | 78% | 81% |
| Mathematics | ||
| AIME 2023 | 77% | — |
| AIME 2024 | 79% | — |
| AIME 2025 | 78% | — |
| HMMT Feb 2023 | 73% | — |
| HMMT Feb 2024 | 75% | — |
| HMMT Feb 2025 | 74% | — |
| BRUMO 2025 | 76% | — |
| MATH-500 | 82% | — |
Kimi K2.5 is ahead overall, 71 to 67. The biggest single separator in this matchup is SuperGPQA, where the scores are 74% and 63.4%.
Qwen3.5-35B-A3B has the edge for knowledge tasks in this comparison, averaging 79.3 versus 62.8. Inside this category, SuperGPQA is the benchmark that creates the most daylight between them.
Qwen3.5-35B-A3B has the edge for coding in this comparison, averaging 72.6 versus 63.2. Inside this category, LiveCodeBench is the benchmark that creates the most daylight between them.
Kimi K2.5 has the edge for reasoning in this comparison, averaging 69.3 versus 59. Inside this category, LongBench v2 is the benchmark that creates the most daylight between them.
Kimi K2.5 has the edge for agentic tasks in this comparison, averaging 57.6 versus 50.5. Inside this category, Terminal-Bench 2.0 is the benchmark that creates the most daylight between them.
Qwen3.5-35B-A3B has the edge for multimodal and grounded tasks in this comparison, averaging 75.1 versus 74.2. Inside this category, MMMU-Pro is the benchmark that creates the most daylight between them.
Kimi K2.5 has the edge for instruction following in this comparison, averaging 94 versus 91.9. Inside this category, IFEval is the benchmark that creates the most daylight between them.
Qwen3.5-35B-A3B has the edge for multilingual tasks in this comparison, averaging 81 versus 79.8. Inside this category, MMLU-ProX is the benchmark that creates the most daylight between them.
Get notified when new models drop, benchmark scores change, or the leaderboard shifts. One email per week.
Free. No spam. Unsubscribe anytime. We only store derived location metadata for consent routing.