Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.
1-bit Bonsai 4B
~44
Winner · 1/8 categoriesDeepSeek V3.1 (Reasoning)
43
3/8 categories1-bit Bonsai 4B· DeepSeek V3.1 (Reasoning)
Pick 1-bit Bonsai 4B if you want the stronger benchmark profile. DeepSeek V3.1 (Reasoning) only becomes the better choice if reasoning is the priority or you need the larger 128K context window.
1-bit Bonsai 4B finishes one point ahead overall, 44 to 43. That is enough to call, but not enough to treat as a blowout. This matchup comes down to a few meaningful edges rather than one model dominating the board.
1-bit Bonsai 4B's sharpest advantage is in mathematics, where it averages 65.8 against 41.1. The single biggest benchmark swing on the page is MuSR, 41.4% to 30%. DeepSeek V3.1 (Reasoning) does hit back in reasoning, so the answer changes if that is the part of the workload you care about most.
DeepSeek V3.1 (Reasoning) is the reasoning model in the pair, while 1-bit Bonsai 4B 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. DeepSeek V3.1 (Reasoning) gives you the larger context window at 128K, compared with 32K for 1-bit Bonsai 4B.
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 | 1-bit Bonsai 4B | DeepSeek V3.1 (Reasoning) |
|---|---|---|
| Agentic | ||
| Terminal-Bench 2.0 | — | 42% |
| BrowseComp | — | 48% |
| OSWorld-Verified | — | 44% |
| Coding | ||
| HumanEval | — | 26% |
| SWE-bench Verified | — | 14% |
| LiveCodeBench | — | 16% |
| SWE-bench Pro | — | 25% |
| Multimodal & Grounded | ||
| MMMU-Pro | — | 37% |
| OfficeQA Pro | — | 47% |
| ReasoningDeepSeek V3.1 (Reasoning) wins | ||
| MuSR | 41.4% | 30% |
| BBH | — | 64% |
| LongBench v2 | — | 57% |
| MRCRv2 | — | 56% |
| KnowledgeDeepSeek V3.1 (Reasoning) wins | ||
| GPQA | 28.7% | 33% |
| MMLU | — | 34% |
| SuperGPQA | — | 31% |
| MMLU-Pro | — | 53% |
| HLE | — | 10% |
| FrontierScience | — | 37% |
| SimpleQA | — | 32% |
| Instruction FollowingDeepSeek V3.1 (Reasoning) wins | ||
| IFEval | 69.6% | 70% |
| Multilingual | ||
| MGSM | — | 64% |
| MMLU-ProX | — | 61% |
| Mathematics1-bit Bonsai 4B wins | ||
| MATH-500 | 65.8% | 62% |
| AIME 2023 | — | 34% |
| AIME 2024 | — | 36% |
| AIME 2025 | — | 35% |
| HMMT Feb 2023 | — | 30% |
| HMMT Feb 2024 | — | 32% |
| HMMT Feb 2025 | — | 31% |
| BRUMO 2025 | — | 33% |
1-bit Bonsai 4B is ahead overall, 44 to 43. The biggest single separator in this matchup is MuSR, where the scores are 41.4% and 30%.
DeepSeek V3.1 (Reasoning) has the edge for knowledge tasks in this comparison, averaging 32.5 versus 28.7. Inside this category, GPQA is the benchmark that creates the most daylight between them.
1-bit Bonsai 4B has the edge for math in this comparison, averaging 65.8 versus 41.1. Inside this category, MATH-500 is the benchmark that creates the most daylight between them.
DeepSeek V3.1 (Reasoning) has the edge for reasoning in this comparison, averaging 49.5 versus 41.4. Inside this category, MuSR is the benchmark that creates the most daylight between them.
DeepSeek V3.1 (Reasoning) has the edge for instruction following in this comparison, averaging 70 versus 69.6. Inside this category, IFEval 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.