Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.
Holo2-4B
Coming soon
0/8 categoriesQwen3 235B 2507 (Reasoning)
52
Winner · 0/8 categoriesHolo2-4B· Qwen3 235B 2507 (Reasoning)
Benchmark data for Holo2-4B and Qwen3 235B 2507 (Reasoning) is coming soon on BenchLM.
BenchLM does not have sourced benchmark coverage for Holo2-4B yet. This comparison is currently limited to metadata such as context window, reasoning mode, and pricing where available.
Holo2-4B has the larger context window at 262K, compared with 128K for Qwen3 235B 2507 (Reasoning).
| Benchmark | Holo2-4B | Qwen3 235B 2507 (Reasoning) |
|---|---|---|
| Agentic | ||
| Terminal-Bench 2.0 | — | 47% |
| BrowseComp | — | 48% |
| OSWorld-Verified | — | 43% |
| Coding | ||
| HumanEval | — | 32% |
| SWE-bench Verified | — | 16% |
| LiveCodeBench | — | 74.1% |
| SWE-bench Pro | — | 29% |
| Multimodal & Grounded | ||
| MMMU-Pro | — | 38% |
| OfficeQA Pro | — | 47% |
| Reasoning | ||
| MuSR | — | 36% |
| BBH | — | 63% |
| LongBench v2 | — | 58% |
| MRCRv2 | — | 58% |
| Knowledge | ||
| MMLU | — | 40% |
| GPQA | — | 81.1% |
| SuperGPQA | — | 64.9% |
| MMLU-Pro | — | 84.4% |
| HLE | — | 6% |
| FrontierScience | — | 42% |
| SimpleQA | — | 38% |
| Instruction Following | ||
| IFEval | — | 87.8% |
| Multilingual | ||
| MGSM | — | 62% |
| MMLU-ProX | — | 81% |
| Mathematics | ||
| AIME 2023 | — | 40% |
| AIME 2024 | — | 42% |
| AIME 2025 | — | 92.3% |
| HMMT Feb 2023 | — | 36% |
| HMMT Feb 2024 | — | 38% |
| HMMT Feb 2025 | — | 37% |
| BRUMO 2025 | — | 39% |
| MATH-500 | — | 60% |
Not fully yet. BenchLM is tracking both models, but the sourced benchmark breakdown for this comparison is still coming soon.
BenchLM only shows category winners and benchmark-level calls when we have sourced results that can be compared fairly. For these models, the public benchmark coverage is not complete enough yet.
Qwen3 235B 2507 (Reasoning): $0.00 input / $0.00 output per 1M tokens Both model pages still include creator, context window, reasoning mode, and other metadata while benchmark coverage fills in.
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.