Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.
1-bit Bonsai 4B
~44
Winner · 2/8 categoriesLlama 4 Behemoth
35
2/8 categories1-bit Bonsai 4B· Llama 4 Behemoth
Pick 1-bit Bonsai 4B if you want the stronger benchmark profile. Llama 4 Behemoth only becomes the better choice if knowledge is the priority.
1-bit Bonsai 4B is clearly ahead on the aggregate, 44 to 35. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
1-bit Bonsai 4B's sharpest advantage is in mathematics, where it averages 65.8 against 51.1. The single biggest benchmark swing on the page is GPQA, 28.7% to 47%. Llama 4 Behemoth does hit back in knowledge, so the answer changes if that is the part of the workload you care about most.
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 | Llama 4 Behemoth |
|---|---|---|
| Agentic | ||
| Terminal-Bench 2.0 | — | 33% |
| BrowseComp | — | 38% |
| OSWorld-Verified | — | 34% |
| Coding | ||
| HumanEval | — | 40% |
| SWE-bench Verified | — | 15% |
| LiveCodeBench | — | 13% |
| SWE-bench Pro | — | 15% |
| Multimodal & Grounded | ||
| MMMU-Pro | — | 60% |
| OfficeQA Pro | — | 49% |
| ReasoningLlama 4 Behemoth wins | ||
| MuSR | 41.4% | 44% |
| BBH | — | 62% |
| LongBench v2 | — | 46% |
| MRCRv2 | — | 46% |
| KnowledgeLlama 4 Behemoth wins | ||
| GPQA | 28.7% | 47% |
| MMLU | — | 48% |
| SuperGPQA | — | 45% |
| MMLU-Pro | — | 54% |
| HLE | — | 3% |
| FrontierScience | — | 43% |
| SimpleQA | — | 46% |
| Instruction Following1-bit Bonsai 4B wins | ||
| IFEval | 69.6% | 68% |
| Multilingual | ||
| MGSM | — | 66% |
| MMLU-ProX | — | 61% |
| Mathematics1-bit Bonsai 4B wins | ||
| MATH-500 | 65.8% | 60% |
| AIME 2023 | — | 48% |
| AIME 2024 | — | 50% |
| AIME 2025 | — | 49% |
| HMMT Feb 2023 | — | 44% |
| HMMT Feb 2024 | — | 46% |
| HMMT Feb 2025 | — | 45% |
| BRUMO 2025 | — | 47% |
1-bit Bonsai 4B is ahead overall, 44 to 35. The biggest single separator in this matchup is GPQA, where the scores are 28.7% and 47%.
Llama 4 Behemoth has the edge for knowledge tasks in this comparison, averaging 37.3 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 51.1. Inside this category, MATH-500 is the benchmark that creates the most daylight between them.
Llama 4 Behemoth has the edge for reasoning in this comparison, averaging 45.5 versus 41.4. Inside this category, MuSR is the benchmark that creates the most daylight between them.
1-bit Bonsai 4B has the edge for instruction following in this comparison, averaging 69.6 versus 68. 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.