Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.
1-bit Bonsai 1.7B
~39
1/8 categoriesDeepSeek V3.1
40
Winner · 3/8 categories1-bit Bonsai 1.7B· DeepSeek V3.1
Pick DeepSeek V3.1 if you want the stronger benchmark profile. 1-bit Bonsai 1.7B only becomes the better choice if reasoning is the priority.
DeepSeek V3.1 finishes one point ahead overall, 40 to 39. 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.
DeepSeek V3.1's sharpest advantage is in knowledge, where it averages 30.3 against 20.7. The single biggest benchmark swing on the page is MATH-500, 34.4% to 59%. 1-bit Bonsai 1.7B does hit back in reasoning, so the answer changes if that is the part of the workload you care about most.
DeepSeek V3.1 gives you the larger context window at 128K, compared with 32K for 1-bit Bonsai 1.7B.
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 1.7B | DeepSeek V3.1 |
|---|---|---|
| Agentic | ||
| Terminal-Bench 2.0 | — | 29% |
| BrowseComp | — | 39% |
| OSWorld-Verified | — | 33% |
| Coding | ||
| HumanEval | — | 25% |
| SWE-bench Verified | — | 13% |
| LiveCodeBench | — | 15% |
| Multimodal & Grounded | ||
| MMMU-Pro | — | 35% |
| OfficeQA Pro | — | 45% |
| Reasoning1-bit Bonsai 1.7B wins | ||
| MuSR | 45.1% | 29% |
| BBH | — | 61% |
| MRCRv2 | — | 48% |
| KnowledgeDeepSeek V3.1 wins | ||
| GPQA | 20.7% | 32% |
| MMLU | — | 33% |
| SuperGPQA | — | 30% |
| MMLU-Pro | — | 53% |
| HLE | — | 2% |
| FrontierScience | — | 37% |
| SimpleQA | — | 31% |
| Instruction FollowingDeepSeek V3.1 wins | ||
| IFEval | 63% | 67% |
| Multilingual | ||
| MGSM | — | 64% |
| MMLU-ProX | — | 59% |
| MathematicsDeepSeek V3.1 wins | ||
| MATH-500 | 34.4% | 59% |
| AIME 2023 | — | 33% |
| AIME 2024 | — | 35% |
| AIME 2025 | — | 34% |
| HMMT Feb 2023 | — | 29% |
| HMMT Feb 2024 | — | 31% |
| HMMT Feb 2025 | — | 30% |
| BRUMO 2025 | — | 32% |
DeepSeek V3.1 is ahead overall, 40 to 39. The biggest single separator in this matchup is MATH-500, where the scores are 34.4% and 59%.
DeepSeek V3.1 has the edge for knowledge tasks in this comparison, averaging 30.3 versus 20.7. Inside this category, GPQA is the benchmark that creates the most daylight between them.
DeepSeek V3.1 has the edge for math in this comparison, averaging 39.6 versus 34.4. Inside this category, MATH-500 is the benchmark that creates the most daylight between them.
1-bit Bonsai 1.7B has the edge for reasoning in this comparison, averaging 45.1 versus 39.6. Inside this category, MuSR is the benchmark that creates the most daylight between them.
DeepSeek V3.1 has the edge for instruction following in this comparison, averaging 67 versus 63. 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.