Claude 4.5 Opus vs Bonsai 2 27B — GGUF
New ComparisonScores are fetched live from the current database on every visit — nothing is cached.
Benchmark Scores Comparison
Claude 4.5 Opus
Bonsai 2 27B — GGUF
Claude 4.5 Opus
Score: 15.0
Bonsai 2 27B — GGUF
27.4B params
Coverage matrix
2 models × 46 benchmarks — hover a benchmark name for its category
Shading = score relative to the best score among the selected models:
Low
Mid
High
Missing (—)
The totals row shows how many of the 46 benchmarks each model has scores for.
| Benchmark | Bonsai 2 27B — GGUF Prism ML | Claude 4.5 Opus Anthropic |
|---|---|---|
| HMMT Feb 25 | — | 92.9 |
| DynaMath | — | 79.7 |
| GPQA Diamond | — | 87 |
| IFBench (prompt loose) | 74 | — |
| LiveCodeBench v6 | 90.07 | 84.8 |
| SWE-bench Pro | — | 57.1 |
| AIME 26 | 95.83 | 95.1 |
| MMStar | — | 73.2 |
| GSM8K | 96.66 | — |
| MBPP+ | 83.07 | — |
| NL2Repo | — | 43.2 |
| MMMU-Pro | 75.49 | 70.6 |
| QwenClawBench | — | 52.3 |
| Claw-Eval Avg | — | 76.6 |
| VideoMMMU | — | 84.4 |
| HMMT Nov 25 | — | 93.3 |
| HumanEval+ | 95.12 | — |
| MMLU-Pro | — | 89.5 |
| SuperGPQA | — | 70.6 |
| IFEval | 91.31 | — |
| CharXiv (RQ) | — | 68.5 |
| MMLU-Redux | 89.09 | 95.6 |
| SWE-bench Multilingual | — | 77.5 |
| MLVU | — | 81.7 |
| OCR Bench v2 | 56.88 | — |
| SWE-bench Verified | — | 80.9 |
| IMOAnswerBench | — | 84 |
| Humanity's Last Exam | — | 30.8 |
| CountBench | — | 90.6 |
| ERQA | — | 46.8 |
| Terminal-Bench 2.0 | — | 59.3 |
| QwenWebBench | — | 1536 |
| MATH-500 | 98.8 | — |
| HMMT Feb 26 | — | 85.3 |
| AIME 2025 | 95 | — |
| CC-OCR | — | 76.9 |
| MuSR | 70.63 | — |
| C-Eval | — | 92.2 |
| SimpleVQA | — | 65.7 |
| VideoMME (w sub.) | — | 77.7 |
| SkillsBench Avg5 | — | 45.3 |
| Claw-Eval Pass^3 | — | 59.6 |
| V-Star | — | 67 |
| BFCL-V3 | 74.92 | — |
| MMMU | — | 80.7 |
| MVBench | — | 67.2 |
| Covered | 14/46 | 36/46 |
Fully comparable benchmarks
Scored by all 2 selected models (4)
| Benchmark | Category | Claude 4.5 Opus | Bonsai 2 27B — GGUF |
|---|---|---|---|
| LiveCodeBench v6 | stem_reasoning | 84.8 | 90.07 |
| AIME 26 | stem_reasoning | 95.1 | 95.83 |
| MMMU-Pro | vision_language | 70.6 | 75.49 |
| MMLU-Redux | knowledge | 95.6 | 89.09 |
Partial coverage benchmarks
Scored by at least one but not all selected models (42) — missing scores show as —
| Benchmark | Category | Claude 4.5 Opus | Bonsai 2 27B — GGUF |
|---|---|---|---|
| HMMT Feb 25 1/2 models | stem_reasoning | 92.9 | — |
| DynaMath 1/2 models | vision_language | 79.7 | — |
| GPQA Diamond 1/2 models | stem_reasoning | 87 | — |
| IFBench (prompt loose) 1/2 models | instruction_following | — | 74 |
| SWE-bench Pro 1/2 models | coding_agent | 57.1 | — |
| MMStar 1/2 models | vision_language | 73.2 | — |
| GSM8K 1/2 models | math | — | 96.66 |
| MBPP+ 1/2 models | code | — | 83.07 |
| NL2Repo 1/2 models | coding_agent | 43.2 | — |
| QwenClawBench 1/2 models | coding_agent | 52.3 | — |
| Claw-Eval Avg 1/2 models | coding_agent | 76.6 | — |
| VideoMMMU 1/2 models | video_understanding | 84.4 | — |
| HMMT Nov 25 1/2 models | stem_reasoning | 93.3 | — |
| HumanEval+ 1/2 models | code | — | 95.12 |
| MMLU-Pro 1/2 models | knowledge | 89.5 | — |
| SuperGPQA 1/2 models | knowledge | 70.6 | — |
| IFEval 1/2 models | instruction_following | — | 91.31 |
| CharXiv (RQ) 1/2 models | document_understanding | 68.5 | — |
| SWE-bench Multilingual 1/2 models | coding_agent | 77.5 | — |
| MLVU 1/2 models | video_understanding | 81.7 | — |
| OCR Bench v2 1/2 models | document_understanding | — | 56.88 |
| SWE-bench Verified 1/2 models | coding_agent | 80.9 | — |
| IMOAnswerBench 1/2 models | stem_reasoning | 84 | — |
| Humanity's Last Exam 1/2 models | stem_reasoning | 30.8 | — |
| CountBench 1/2 models | spatial_intelligence | 90.6 | — |
| ERQA 1/2 models | spatial_intelligence | 46.8 | — |
| Terminal-Bench 2.0 1/2 models | coding_agent | 59.3 | — |
| QwenWebBench 1/2 models | coding_agent | 1536 | — |
| MATH-500 1/2 models | math | — | 98.8 |
| HMMT Feb 26 1/2 models | stem_reasoning | 85.3 | — |
| AIME 2025 1/2 models | stem_reasoning | — | 95 |
| CC-OCR 1/2 models | document_understanding | 76.9 | — |
| MuSR 1/2 models | reasoning | — | 70.63 |
| C-Eval 1/2 models | knowledge | 92.2 | — |
| SimpleVQA 1/2 models | vision_language | 65.7 | — |
| VideoMME (w sub.) 1/2 models | video_understanding | 77.7 | — |
| SkillsBench Avg5 1/2 models | coding_agent | 45.3 | — |
| Claw-Eval Pass^3 1/2 models | coding_agent | 59.6 | — |
| V-Star 1/2 models | vision_language | 67 | — |
| BFCL-V3 1/2 models | general_agent | — | 74.92 |
| MMMU 1/2 models | vision_language | 80.7 | — |
| MVBench 1/2 models | video_understanding | 67.2 | — |