Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.
Phi-4
40
0/8 categoriesQwen3.5-35B-A3B
67
Winner · 6/8 categoriesPhi-4· Qwen3.5-35B-A3B
Pick Qwen3.5-35B-A3B if you want the stronger benchmark profile. Phi-4 only becomes the better choice if you would rather avoid the extra latency and token burn of a reasoning model.
Qwen3.5-35B-A3B is clearly ahead on the aggregate, 67 to 40. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
Qwen3.5-35B-A3B's sharpest advantage is in multimodal & grounded, where it averages 75.1 against 46.8. The single biggest benchmark swing on the page is LongBench v2, 30% to 59%.
Qwen3.5-35B-A3B is the reasoning model in the pair, while Phi-4 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 16K for Phi-4.
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 | Phi-4 | Qwen3.5-35B-A3B |
|---|---|---|
| AgenticQwen3.5-35B-A3B wins | ||
| Terminal-Bench 2.0 | 44% | 40.5% |
| BrowseComp | 35% | 61% |
| OSWorld-Verified | 34% | 54.5% |
| tau2-bench | — | 81.2% |
| CodingQwen3.5-35B-A3B wins | ||
| HumanEval | 82.6% | — |
| SWE-bench Pro | 55% | — |
| SWE-bench Verified | — | 69.2% |
| LiveCodeBench | — | 74.6% |
| Multimodal & GroundedQwen3.5-35B-A3B wins | ||
| MMMU-Pro | 54% | 75.1% |
| OfficeQA Pro | 38% | — |
| ReasoningQwen3.5-35B-A3B wins | ||
| LongBench v2 | 30% | 59% |
| MRCRv2 | 33% | — |
| KnowledgeQwen3.5-35B-A3B wins | ||
| MMLU | 84.8% | — |
| GPQA | 56.1% | 84.2% |
| FrontierScience | 52% | — |
| MMLU-Pro | — | 85.3% |
| SuperGPQA | — | 63.4% |
| Instruction Following | ||
| IFEval | — | 91.9% |
| MultilingualQwen3.5-35B-A3B wins | ||
| MGSM | 80.6% | — |
| MMLU-ProX | 60% | 81% |
| Mathematics | ||
| MATH-500 | 94.6% | — |
Qwen3.5-35B-A3B is ahead overall, 67 to 40. The biggest single separator in this matchup is LongBench v2, where the scores are 30% and 59%.
Qwen3.5-35B-A3B has the edge for knowledge tasks in this comparison, averaging 79.3 versus 53.6. Inside this category, GPQA 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 55. Phi-4 stays close enough that the answer can still flip depending on your workload.
Qwen3.5-35B-A3B has the edge for reasoning in this comparison, averaging 59 versus 31.4. Inside this category, LongBench v2 is the benchmark that creates the most daylight between them.
Qwen3.5-35B-A3B has the edge for agentic tasks in this comparison, averaging 50.5 versus 38.3. Inside this category, BrowseComp 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 46.8. Inside this category, MMMU-Pro 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 67.2. 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.