Claude-Sonnet-4.5 vs Claude 4.5 Opus vs Claude Opus 4.8 vs Claude Opus 4.6 vs Claude Sonnet 4.6 vs Claude Opus 4.7 vs DeepSeek V4 Pro vs DeepSeek-V3.2 vs Claude-Opus-4.5 vs Claude-Opus-4.6 (Max) vs Claude Opus 5 vs Clef vs Bonsai 2 27B — GGUF vs Clef-Flash vs d1-2B
New ComparisonScores are fetched live from the current database on every visit — nothing is cached.
Winner
Claude Opus 5
Claude Opus 5 scores higher overall across category strength profiles (60 vs 52 for Claude Opus 4.8); only 0 benchmarks are directly comparable.
Claude Opus 5
60.4
Claude Opus 4.8
52.4
Claude 4.5 Opus
49
Claude Opus 4.7
47.1
DeepSeek V4 Pro
46.7
Claude Opus 4.6
44.7
Claude Sonnet 4.6
39.1
d1-2B
36.1
Bonsai 2 27B — GGUF
33.3
Clef-Flash
32.7
Claude-Sonnet-4.5
31.4
Claude-Opus-4.6 (Max)
31
Clef
30.9
DeepSeek-V3.2
28.3
Claude-Opus-4.5
24.2
Only 0 of 224 benchmarks are scored by every selected model — the verdict leans on category strength profiles and is indicative.
Benchmark Scores Comparison
Claude-Sonnet-4.5
Claude 4.5 Opus
Claude Opus 4.8
Claude Opus 4.6
Claude Sonnet 4.6
Claude Opus 4.7
DeepSeek V4 Pro
DeepSeek-V3.2
Claude-Opus-4.5
Claude-Opus-4.6 (Max)
Claude Opus 5
Clef
Bonsai 2 27B — GGUF
Clef-Flash
d1-2B
Claude-Sonnet-4.5
Score: 15.0
Claude 4.5 Opus
Score: 15.0
Claude Opus 4.8
Score: 15.0
Claude Opus 4.6
Score: 15.0
Claude Sonnet 4.6
Score: 15.0
Claude Opus 4.7
Score: 15.0
DeepSeek V4 Pro
Score: 15.0
DeepSeek-V3.2
Score: 15.0
Claude-Opus-4.5
Score: 15.0
Claude-Opus-4.6 (Max)
Score: 40.0
Claude Opus 5
Score: 40.0
Clef
27.0B params
Bonsai 2 27B — GGUF
27.4B params
Clef-Flash
9.0B params
d1-2B
1.9B params
Coverage matrix
15 models × 224 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 224 benchmarks each model has scores for.
| Benchmark | Claude-Opus-4.6 (Max) Anthropic | Bonsai 2 27B — GGUF Prism ML | Clef Cloudflare | Clef-Flash Cloudflare | d1-2B LiquidAI | Claude-Sonnet-4.5 Anthropic | Claude 4.5 Opus Anthropic | Claude Opus 4.8 Anthropic | Claude Opus 4.6 Anthropic | Claude Sonnet 4.6 Anthropic | Claude Opus 4.7 Anthropic | DeepSeek V4 Pro DeepSeek | DeepSeek-V3.2 DeepSeek | Claude-Opus-4.5 Anthropic | Claude Opus 5 Anthropic |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| SVG-Bench | — | — | — | — | — | — | — | — | — | 64.1 | 62.3 | 61.7 | — | — | — |
| Amazon ESCI | — | — | 57.5 | 57.4 | — | — | — | — | — | — | — | — | — | — | — |
| BPoMP | — | — | 96.9 | 95.4 | — | — | — | — | — | — | — | — | — | — | — |
| HMMT Feb 25 | — | — | — | — | — | — | 92.9 | — | — | — | — | — | — | — | — |
| ForecastBench | — | — | 13.9 | 10.6 | — | — | — | — | — | — | — | — | — | — | — |
| DynaMath | — | — | — | — | — | 78.8 | 79.7 | — | — | — | — | — | — | — | — |
| GPQA Diamond | 91.3 | — | 48 | 51 | — | — | 87 | 92 | 91.3 | — | — | 92.8 | — | — | 93.7 |
| IFBench (prompt loose) | — | 74 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| MathVista (mini) | — | — | — | — | — | 79.8 | — | — | — | — | — | — | — | — | — |
| MMBench EN-DEV v1.1 | — | — | — | — | 87.9 | 88.3 | — | — | — | — | — | — | — | — | — |
| ArXivMath | — | — | — | — | — | — | — | — | — | — | — | 62.1 | — | — | 71.5 |
| LiveCodeBench v6 | 88.8 | 90.07 | — | — | — | — | 84.8 | — | 88.8 | — | — | — | — | — | — |
| WorkflowEvals | — | — | — | — | 81.6 | — | — | — | — | — | — | — | — | — | — |
| Hy-LifeSearch (Internal) | — | — | — | — | — | — | — | — | — | — | — | 46.9 | — | — | 56.1 |
| FinEntity | — | — | 96.2 | 97.1 | — | — | — | — | — | — | — | — | — | — | — |
| ChessBench | — | — | 24.7 | 23 | — | — | — | — | — | — | — | — | — | — | — |
| MRCR v2 256K (8-needle) | — | — | — | — | — | — | — | 83.2 | — | — | — | — | — | — | — |
| Fast Decisions | — | — | — | — | 58.8 | — | — | — | — | — | — | — | — | — | — |
| Automation-Bench | — | — | — | — | — | — | — | 27.2 | — | — | — | 30.4 | — | — | 48.7 |
| SWE-bench Pro | 53.4 | — | — | — | — | 44 | 57.1 | 69.2 | 53.4 | — | 64.3 | 55.4 | 40.9 | 47 | 79.2 |
| NLI4CT | — | — | 82.9 | 78.6 | — | — | — | — | — | — | — | — | — | — | — |
| RealWorldQA | 73.9 | — | — | — | — | 70.3 | — | — | — | — | — | — | — | — | — |
| BioMysteryBench | — | — | — | — | — | — | — | — | — | — | — | 61.6 | — | — | 72.1 |
| Home appliance simulator | — | — | 83 | 97.7 | — | — | — | — | — | — | — | — | — | — | — |
| PRBench-Finance | — | — | — | — | — | — | — | 51.9 | — | — | — | — | — | — | — |
| EmbSpatialBench | — | — | — | — | — | 71.8 | — | — | — | — | — | — | — | — | — |
| MathVision | 65.5 | — | — | — | — | 71.1 | — | — | 71.2 | — | — | — | — | — | — |
| AIME 26 | — | 95.83 | — | — | — | — | 95.1 | 95.7 | 96.7 | — | — | — | — | — | — |
| LiveSQLBench | — | — | — | — | — | — | — | — | — | — | 41 | — | — | — | — |
| AndroidWorld | 62 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| ACOS | — | — | 33.3 | 25.9 | — | — | — | — | — | — | — | — | — | — | — |
| AgentWorldBench Terminal | — | — | — | — | — | — | — | 59.18 | 57.51 | 56.98 | — | — | — | — | — |
| ContractNLI | — | — | 81.4 | 84.3 | — | — | — | — | — | — | — | — | — | — | — |
| Agents' Last Exam | — | — | — | — | — | — | — | 27 | — | — | — | 21.9 | — | — | 25.1 |
| Toolathlon Verified | — | — | — | — | — | — | — | 76.2 | 47.2 | — | — | 74.1 | — | — | 76.5 |
| CLINC150+OOS | — | — | 97.4 | 66.8 | — | — | — | — | — | — | — | — | — | — | — |
| Hy-Backend 2.0 (Internal) | — | — | — | — | — | — | — | — | — | — | — | 33.8 | — | — | 40.3 |
| PLawBench | — | — | — | — | — | — | — | 69.6 | — | — | — | — | — | — | — |
| ScreenSpot Pro | — | — | — | — | — | 36.2 | — | — | — | — | — | — | — | — | — |
| OJBench | — | — | — | — | — | — | — | — | 60.3 | — | — | — | — | — | — |
| Hy-SWE Max Verified (Internal) | — | — | — | — | — | — | — | — | — | — | — | 65.7 | — | — | 70.1 |
| AgentWorldBench Overall | — | — | — | — | — | — | — | 56.59 | 57.8 | 56.04 | — | — | — | — | — |
| Hy-BrowseComp-Pro2 (Internal) | — | — | — | — | — | — | — | — | — | — | — | 46.5 | — | — | 61.3 |
| ARC-c | — | — | 97.7 | 98.3 | — | — | — | — | — | — | — | — | — | — | — |
| QwenReactBench | — | — | — | — | — | — | — | 1694 | — | — | — | — | — | — | — |
| MMStar | — | — | — | — | 67.5 | 73.8 | 73.2 | — | — | — | — | — | — | — | — |
| GSM8K | — | 96.66 | 80.8 | 67.3 | — | — | — | — | — | — | — | — | — | — | — |
| ToolRet | — | — | 69.2 | 66.4 | — | — | — | — | — | — | — | — | — | — | — |
| Decision Bench (decisionbench.ai) | — | — | — | — | 85.2 | — | — | — | — | — | — | — | — | — | — |
| GDPVal | — | — | — | — | — | — | — | — | — | 75.7 | 79.8 | 70.3 | — | — | — |
| OCRBench | — | — | — | — | — | 76.6 | — | — | — | — | — | — | — | — | — |
| OfficeQA Pro | — | — | — | — | — | — | — | — | — | — | 43.6 | 65.4 | — | — | 66.9 |
| PMC-VQA | — | — | — | — | — | 55.9 | — | — | — | — | — | — | — | — | — |
| QwenSVGBench | — | — | — | — | — | — | — | 1648 | — | — | — | — | — | — | — |
| DeepSearch QA | — | — | — | — | — | — | — | — | 91.3 | — | — | — | — | — | — |
| RouterBench | — | — | 79.7 | 79.9 | — | — | — | — | — | — | — | — | — | — | — |
| SWE Atlas - TW | — | — | — | — | — | — | — | — | — | 31.8 | 38.2 | 45.6 | — | — | 68.7 |
| CL-bench | — | — | — | — | — | — | — | — | — | — | 22.9 | — | — | — | — |
| MBPP+ | — | 83.07 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| LOCA-Bench (256k) | — | — | — | — | — | — | — | — | — | — | 57 | — | — | — | 53.7 |
| KernelBench Hard | — | — | — | — | — | — | — | — | — | — | 30.7 | — | — | — | — |
| AgentWorldBench Web | — | — | — | — | — | — | — | 54.66 | 51.42 | 50.78 | — | — | — | — | — |
| SWE Atlas - QnA | — | — | — | — | — | — | — | — | — | 31.2 | 45.2 | 53.4 | — | — | 57.5 |
| AI2D_TEST | — | — | — | — | 76.7 | 87 | — | — | — | — | — | — | — | — | — |
| LingoQA | — | — | — | — | — | 12.8 | — | — | — | — | — | — | — | — | — |
| IMO 2025 | — | — | — | — | — | — | — | — | — | — | 17.4 | — | — | — | — |
| NL2Repo | — | — | — | — | — | — | 43.2 | 69.7 | — | — | — | 61.5 | — | — | 75.3 |
| PRBench-Legal | — | — | — | — | — | — | — | 52.7 | — | — | — | — | — | — | — |
| MMMU-Pro | — | 75.49 | — | — | 26.4 | 68.4 | 70.6 | — | 73.9 | 74.5 | 77 | — | — | — | — |
| Program Bench | — | — | — | — | — | — | — | 63.8 | — | — | — | — | — | — | — |
| MCPMark-Verified | — | — | — | — | — | — | — | 76.4 | — | — | — | — | — | — | — |
| PostTrainBench | — | — | — | — | — | — | — | 37.2 | — | — | 42.4 | 24.5 | — | — | 35 |
| SGD/SGD-X | — | — | 43.8 | 34.2 | — | — | — | — | — | — | — | — | — | — | — |
| QwenClawBench | — | — | — | — | — | — | 52.3 | — | — | — | — | — | — | — | — |
| BabyVision | — | — | — | — | — | 18.6 | — | — | 14.8 | — | — | — | — | — | 94.1 |
| SLAKE | — | — | — | — | — | 73.6 | — | — | — | — | — | — | — | — | — |
| MMMLU | — | — | — | — | — | — | — | — | — | — | — | — | 87.9 | — | — |
| Claw-Eval Avg | 54.7 | — | — | — | — | — | 76.6 | — | 82.4 | 68.3 | 71.6 | 58.4 | — | — | — |
| ANLI | — | — | 69.8 | 59.1 | — | — | — | — | — | — | — | — | — | — | — |
| VideoMMMU | — | — | — | — | — | 77.6 | 84.4 | — | — | — | 83 | — | — | — | — |
| HMMT Nov 25 | — | — | — | — | — | — | 93.3 | 96.5 | — | — | — | — | — | — | — |
| Apex-Agents | — | — | — | — | — | — | — | — | 33 | 26.2 | 37.2 | 32.4 | — | — | 41.8 |
| HumanEval+ | — | 95.12 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| ARC-e | — | — | 99 | 99.5 | — | — | — | — | — | — | — | — | — | — | — |
| VisualWebBench | — | — | — | — | 96.2 | — | — | — | — | — | — | — | — | — | — |
| MMRB2 | — | — | — | — | 54.2 | — | — | — | — | — | — | — | — | — | — |
| WorkSpaceBench | — | — | — | — | — | — | — | 66.8 | — | — | — | 65.4 | — | — | 75 |
| MMLU-Pro | — | — | 65.9 | 65.3 | — | — | 89.5 | — | — | — | — | — | 65.5 | — | — |
| NaturalBench | — | — | — | — | 81.5 | — | — | — | — | — | — | — | — | — | — |
| BankerToolBench | — | — | — | — | — | — | — | — | — | — | 81.3 | 73.1 | — | — | 81.9 |
| SuperGPQA | — | — | — | — | — | — | 70.6 | — | — | — | — | — | 45 | — | — |
| E-Bench-Code (Internal) | — | — | — | — | — | — | — | — | — | — | — | 64.3 | — | — | 82.7 |
| OSWorld-Verified | — | — | — | — | — | 61.4 | — | — | 72.7 | 72.5 | 82.8 | — | — | — | — |
| TIR-Bench | — | — | — | — | — | 27.6 | — | — | — | — | — | — | — | — | — |
| AgentWorldBench Search | — | — | — | — | — | — | — | 35.14 | 29.3 | 28.79 | — | — | — | — | — |
| POP909-CL | — | — | 15.8 | 1.6 | — | — | — | — | — | — | — | — | — | — | — |
| CV-Bench | — | — | — | — | 86.1 | — | — | — | — | — | — | — | — | — | — |
| HellaSwag | — | — | 98.2 | 98.6 | — | — | — | — | — | — | — | — | — | — | — |
| IFEval | — | 91.31 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| HLE (with tools) | 40 | — | — | — | — | — | — | 57.9 | 53 | — | — | 60 | — | — | 60.9 |
| LVBench | 63 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| CharXiv (RQ) | 66 | — | — | — | — | 67.2 | 68.5 | — | 69.1 | — | — | — | — | — | 83.7 |
| Kimi Claw 24/7 Bench | — | — | — | — | — | — | — | 50.4 | — | — | — | — | — | — | — |
| POPE | — | — | — | — | 88.7 | — | — | — | — | — | — | — | — | — | — |
| CLadder | — | — | 94 | 97.7 | — | — | — | — | — | — | — | — | — | — | — |
| HoVer | — | — | 65.2 | 61.2 | — | — | — | — | — | — | — | — | — | — | — |
| FrontierSWE | — | — | — | — | — | — | — | 70 | — | — | — | — | — | — | — |
| GDP.pdf | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 21.6 |
| cfcolor | — | — | 66 | 65.8 | — | — | — | — | — | — | — | — | — | — | — |
| CritPt (no tools) | — | — | — | — | — | — | — | 20.9 | — | — | — | 18 | — | — | 29.1 |
| BrowseComp | — | — | — | — | — | — | — | — | 83.7 | 74.7 | 79.3 | 83.4 | — | — | — |
| MedXpertQA-MM | — | — | — | — | — | 54 | — | — | — | — | — | — | — | — | — |
| MMLU-Redux | — | 89.09 | — | — | — | — | 95.6 | — | — | — | — | — | 87.5 | — | — |
| SATA-Bench | — | — | 33.8 | 36.7 | — | — | — | — | — | — | — | — | — | — | — |
| NL2Repo-Bench | 47.6 | — | — | — | — | — | — | 69.4 | — | — | 56.3 | 35.5 | — | — | 75.3 |
| BRIGHT | — | — | 45.9 | 39.3 | — | — | — | — | — | — | — | — | — | — | — |
| DRACO | — | — | — | — | — | — | — | — | — | 75.8 | 77.7 | 77.3 | — | — | 88.6 |
| SkillsBench | — | — | — | — | — | — | — | 65.1 | — | — | — | — | — | — | — |
| CoWorkBench | 68.2 | — | — | — | — | — | — | 72.3 | — | — | — | — | — | — | — |
| ImajevBench | — | — | — | — | 75.5 | — | — | — | — | — | — | — | — | — | — |
| MCPMark | — | — | — | — | — | — | — | — | 56.7 | — | — | — | — | — | — |
| SWE-bench Multilingual | 77.5 | — | — | — | — | — | 77.5 | — | 77.8 | — | — | 77.3 | 62.3 | — | 89.5 |
| MMLU | — | — | 90.3 | 91.8 | — | — | — | — | — | — | — | — | 87.8 | — | — |
| MME | — | — | — | — | 85.4 | — | — | — | — | — | — | — | — | — | — |
| VAST | — | — | 59.5 | 49.6 | — | — | — | — | — | — | — | — | — | — | — |
| MMVU | — | — | — | — | — | 70.6 | — | — | — | — | — | — | — | — | — |
| PhishNChips | — | — | 79.6 | 75 | — | — | — | — | — | — | — | — | — | — | — |
| Decision Index (public suite) | — | — | — | — | 51.1 | — | — | — | — | — | — | — | — | — | — |
| PaperBench | — | — | — | — | — | — | — | 80.3 | — | — | 58.5 | — | — | — | — |
| YC-Bench | — | — | — | — | — | — | — | — | — | 100000 | 2200000 | 500000 | — | — | — |
| MLVU | — | — | — | — | — | 72.8 | 81.7 | — | — | — | — | — | — | — | — |
| OCR Bench v2 | — | 56.88 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| JevBench | — | — | — | — | 77.9 | — | — | — | — | — | — | — | — | — | — |
| API-Bank | — | — | 91.9 | 93.1 | — | — | — | — | — | — | — | — | — | — | — |
| Hy-FinAgentBench (Internal) | — | — | — | — | — | — | — | — | — | — | — | 78.5 | — | — | 82 |
| IFBench | 62.5 | — | — | — | — | — | — | 62.2 | — | — | — | — | — | — | — |
| SWE-bench Verified | — | — | — | — | — | — | 80.9 | — | 80.8 | 79.6 | 87.6 | 80.6 | 70.2 | — | — |
| IMOAnswerBench | — | — | — | — | — | — | 84 | 83.5 | 75.3 | — | — | — | — | — | — |
| Humanity's Last Exam | — | — | — | — | — | — | 30.8 | 45.7 | 40 | — | — | 42.7 | — | — | 54.9 |
| CountBench | — | — | — | — | — | 90 | 90.6 | — | — | — | — | — | — | — | — |
| VL-RewardBench | — | — | — | — | 69.3 | — | — | — | — | — | — | — | — | — | — |
| CountQA | — | — | — | — | 51.4 | — | — | — | — | — | — | — | — | — | — |
| ERQA | 40.8 | — | — | — | 45.8 | 45 | 46.8 | — | — | — | — | — | — | — | — |
| VideoMME (w/o sub.) | — | — | — | — | — | 75.3 | — | — | — | — | — | — | — | — | — |
| Terminal-Bench 2.0 | — | — | — | — | — | — | 59.3 | — | 65.4 | — | — | — | 39.3 | — | — |
| VisualPuzzles | — | — | — | — | 29.9 | — | — | — | — | — | — | — | — | — | — |
| CRUXEval | — | — | 86.7 | 86.1 | — | — | — | — | — | — | — | — | — | — | — |
| BIG-Bench Hard | — | — | 73.7 | 68.9 | — | — | — | — | — | — | — | — | — | — | — |
| QwenSWEBench | — | — | — | — | — | — | — | 84 | — | — | — | — | — | — | — |
| SpreadSheetBench-v1 | — | — | — | — | — | — | — | — | — | — | 88.5 | 84.9 | — | — | — |
| QwenWebBench | — | — | — | — | — | — | 1536 | — | — | — | — | — | — | — | — |
| VlmsAreBlind | — | — | — | — | — | 85.5 | — | — | — | — | — | — | — | — | — |
| WinoGrande | — | — | 93.5 | 97.5 | — | — | — | — | — | — | — | — | — | — | — |
| When2Call MCQ | — | — | 72.4 | 65.6 | — | — | — | — | — | — | — | — | — | — | — |
| SUPERChem | — | — | — | — | — | — | — | — | — | — | — | 62 | — | — | 76.7 |
| VIBE-V2 | — | — | — | — | — | — | — | — | — | 42.8 | 55.9 | 45 | — | — | — |
| MATH-500 | — | 98.8 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| HorizonMath (pass@4) | — | — | — | — | — | — | — | — | — | — | — | 4.42 | — | — | 5.3 |
| Terminal-Bench 2.1 (Terminus-2) | — | — | — | — | — | — | — | 85 | — | — | — | 87.9 | — | — | 86.7 |
| RAGTruth | — | — | 79.4 | 35.6 | — | — | — | — | — | — | — | — | — | — | — |
| DeepSWE | — | — | — | — | — | — | — | — | — | — | — | 62.7 | — | — | 68.8 |
| BrokenArXiv | — | — | — | — | — | — | — | — | — | — | — | 43.1 | — | — | 77.7 |
| OmniDocBench 1.5 | — | — | — | — | — | 85.8 | — | — | — | 86.9 | 89.3 | — | — | — | — |
| HallusionBench | — | — | — | — | 72.1 | 59.9 | — | — | — | — | — | — | — | — | — |
| OCRBench | — | — | — | — | — | 76.6 | — | — | — | — | — | — | — | — | — |
| HMMT Feb 26 | — | — | — | — | — | — | 85.3 | 96.7 | 96.2 | — | — | — | — | — | — |
| BANKING77 | — | — | 94.2 | 90.9 | — | — | — | — | — | — | — | — | — | — | — |
| AgentWorldBench OS | — | — | — | — | — | — | — | 66.62 | 70.2 | 63.17 | — | — | — | — | — |
| BLINK | — | — | — | — | 66.2 | — | — | — | — | — | — | — | — | — | — |
| MMVP | — | — | — | — | 82.7 | — | — | — | — | — | — | — | — | — | — |
| AgentWorldBench MCP | — | — | — | — | — | — | — | 54.93 | 69.9 | 70 | — | — | — | — | — |
| Hy-FinmodelBench v2 (Internal) | — | — | — | — | — | — | — | — | — | — | — | 51.3 | — | — | 66 |
| MathArena Apex 2025 | — | — | — | — | — | — | — | — | — | — | — | 66.3 | — | — | 91.4 |
| Terminal Bench 2.1 | — | — | — | — | — | — | — | 84.6 | — | — | 66.1 | 59.6 | — | — | — |
| AIME 2025 | — | 95 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| AgentWorldBench SWE | — | — | — | — | — | — | — | 64.1 | 64.55 | 64.52 | — | — | — | — | — |
| CC-OCR | — | — | — | — | — | 68.1 | 76.9 | — | — | — | — | — | — | — | — |
| JobBench | 36.6 | — | — | — | — | — | — | 48.4 | — | — | — | 54.1 | — | — | 68 |
| $OneMillion-Bench | — | — | — | — | — | — | — | 41.8 | — | — | — | 62 | — | — | 68.1 |
| Kimi Code Bench V2 | — | — | — | — | — | — | — | 67.4 | — | — | — | — | — | — | — |
| SciCode (subtask) | — | — | — | — | — | — | — | — | 51.9 | — | — | — | — | — | — |
| USAMO 2026 | — | — | — | — | — | — | — | — | — | — | 52.8 | — | — | — | — |
| Tool Decathlon | — | — | — | — | — | — | — | 59.9 | — | — | — | — | — | — | — |
| QwenQoderBench | — | — | — | — | — | — | — | 62.7 | — | — | — | — | — | — | — |
| MuSR | — | 70.63 | 83.5 | 86 | — | — | — | — | — | — | — | — | — | — | — |
| OSWorld 2.0 (Binary) | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 70.6 |
| Aider | — | — | — | — | — | — | — | — | — | — | — | — | 69.9 | — | — |
| ZEROBench_sub | — | — | — | — | — | 26.3 | — | — | — | — | — | — | — | — | — |
| C-Eval | — | — | — | — | — | — | 92.2 | — | — | — | — | — | 90.4 | — | — |
| Humicroedit | — | — | 66.7 | 75.1 | — | — | — | — | — | — | — | — | — | — | — |
| WideSearch | — | — | — | — | — | — | — | 72.9 | — | — | — | 81.8 | — | — | 84 |
| SimpleVQA | — | — | — | — | — | 57.6 | 65.7 | — | — | — | — | — | — | — | — |
| VideoMME (w sub.) | — | — | — | — | — | 81.1 | 77.7 | — | — | — | — | — | — | — | — |
| SkillsBench Avg5 | — | — | — | — | — | — | 45.3 | — | — | — | — | 65 | — | — | 63.7 |
| New Yorker | — | — | 69.5 | 66.1 | — | — | — | — | — | — | — | — | — | — | — |
| Habermas Machine | — | — | 68.7 | 71.8 | — | — | — | — | — | — | — | — | — | — | — |
| HealthBench | — | — | — | — | — | — | — | 52.4 | — | — | — | — | — | — | — |
| Claw-Eval Pass^3 | 52.5 | — | — | — | — | — | 59.6 | — | 70.4 | — | — | — | — | — | — |
| E-Bench (Internal) | — | — | — | — | — | — | — | — | — | — | — | 61.3 | — | — | 77.8 |
| ProgramBench | — | — | — | — | — | — | — | 71.9 | — | — | — | 15.5 | — | — | 39.5 |
| LongBench v2 | — | — | — | — | — | — | — | 69.1 | — | — | — | — | 40.2 | — | — |
| MCP-Atlas | — | — | — | — | — | — | — | 77.8 | — | 61.3 | 77 | 73.6 | — | — | 85.7 |
| AndroidBench | — | — | — | — | — | — | — | 69.8 | — | — | — | — | — | — | — |
| SWE-Marathon | — | — | — | — | — | — | — | 26 | — | — | — | 19 | — | — | 50 |
| Terminal-Bench 2.1 (Best Reported Harness) | — | — | — | — | — | — | — | 78.9 | — | — | — | — | — | — | — |
| Cybergym | — | — | — | — | — | — | — | — | — | — | — | 83.3 | — | — | — |
| Terminal-Bench 4.0 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 51.8 |
| DecisionBench (Hanno Labs) | — | — | — | — | 77.1 | — | — | — | — | — | — | — | — | — | — |
| SWE Atlas - RF | — | — | — | — | — | — | — | — | — | — | — | 48.6 | — | — | 60 |
| V-Star | — | — | — | — | 76.4 | 58.6 | 67 | — | 86.4 | — | — | — | — | — | — |
| AgentWorldBench Android | — | — | — | — | — | — | — | 61.5 | 61.74 | 58.03 | — | — | — | — | — |
| DeepSWE 1.1 | — | — | — | — | — | — | — | 59 | — | — | — | — | — | — | 74 |
| Typed Decisions | — | — | — | — | 61.3 | — | — | — | — | — | — | — | — | — | — |
| Hy-CompanyBench V2 (Internal) | — | — | — | — | — | — | — | — | — | — | — | 64.4 | — | — | 72.7 |
| Harbor-Index | — | — | — | — | — | — | — | — | — | — | — | 36.9 | — | — | 56.9 |
| SWE-efficiency | — | — | — | — | — | — | — | — | — | — | 42.2 | — | — | — | — |
| SWE Refactor Bench | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 31 |
| MLS-Bench-Lite | — | — | — | — | — | — | — | 42.8 | — | — | — | — | — | — | — |
| BFCL-V3 | — | 74.92 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| RefSpatialBench | — | — | — | — | — | 2.2 | — | — | — | — | — | — | — | — | — |
| MMMU | — | — | — | — | — | 79.6 | 80.7 | — | — | — | — | — | — | — | — |
| GDPVal-AA v2 | — | — | — | — | — | — | — | 1582 | — | — | — | 1580 | — | — | 1831 |
| MVBench | — | — | — | — | — | — | 67.2 | — | — | — | — | — | — | — | — |
| ZEROBench | — | — | — | — | — | 4 | — | — | — | — | — | — | — | — | — |
| Covered | 17/224 | 14/224 | 40/224 | 40/224 | 26/224 | 38/224 | 36/224 | 57/224 | 34/224 | 23/224 | 31/224 | 56/224 | 12/224 | 1/224 | 54/224 |
Scorecard
Overall 0–100 score — category strengths (50%), shared-benchmark wins (30%), coverage (20%)
| Model | Overall score | Shared benchmark wins | Coverage | Strongest at | Weakest at |
|---|---|---|---|---|---|
| Claude Opus 5 Winner | 60.4 | — | 54 of 224 (24.0%) | STEM & Reasoning | Long Context |
| Claude Opus 4.8 | 52.4 | — | 57 of 224 (25.0%) | Long Context | General Capabilities |
| Claude 4.5 Opus | 49 | — | 36 of 224 (16.0%) | Spatial Intelligence | Document Understanding |
| Claude Opus 4.7 | 47.1 | — | 31 of 224 (14.0%) | Long Context | Vision & Language |
| DeepSeek V4 Pro | 46.7 | — | 56 of 224 (25.0%) | Agentic Coding | Finance |
| Claude Opus 4.6 | 44.7 | — | 34 of 224 (15.0%) | STEM & Reasoning | Document Understanding |
| Claude Sonnet 4.6 | 39.1 | — | 23 of 224 (10.0%) | Document Understanding | Agentic Coding |
| d1-2B | 36.1 | — | 26 of 224 (12.0%) | Spatial Intelligence | Document Understanding |
| Bonsai 2 27B — GGUF | 33.3 | — | 14 of 224 (6.0%) | Math | Reasoning |
| Clef-Flash | 32.7 | — | 40 of 224 (18.0%) | Reasoning | Classification |
| Claude-Sonnet-4.5 | 31.4 | — | 38 of 224 (17.0%) | Video Understanding | General Agents |
| Claude-Opus-4.6 (Max) | 31 | — | 17 of 224 (8.0%) | Instruction Following | Document Understanding |
| Clef | 30.9 | — | 40 of 224 (18.0%) | Safety | General Agents |
| DeepSeek-V3.2 | 28.3 | — | 12 of 224 (5.0%) | Knowledge | Long Context |
| Claude-Opus-4.5 | 24.2 | — | 1 of 224 (0.0%) | Agentic Coding | Agentic Coding |
Category strengths
Where each model stands, per benchmark category — score relative to the best model, adjusted for benchmark coverage. This also ranks models when they share few benchmarks.
| Category | Claude Opus 5 | Claude Opus 4.8 | Claude 4.5 Opus | Claude Opus 4.7 | DeepSeek V4 Pro | Claude Opus 4.6 | Claude Sonnet 4.6 | d1-2B | Bonsai 2 27B — GGUF | Clef-Flash | Claude-Sonnet-4.5 | Claude-Opus-4.6 (Max) | Clef | DeepSeek-V3.2 | Claude-Opus-4.5 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Agentic Coding 45 benchmarks | 82.5 | 63.4 | 53.3 | 59.4 | 61.1 | 59.5 | 45.2 | — | — | — | 31.5 | 41.7 | — | 34.3 | 33.7 |
| General Agents 42 benchmarks | 76.3 | 62.7 | — | 59.9 | 54.8 | 59.3 | 51 | — | — | 59.1 | 30.8 | 33.8 | 0 | — | — |
| Vision & Language 28 benchmarks | 58.5 | — | 64.9 | 58.5 | — | 51 | 55.6 | 47.2 | 56.7 | 0 | 54.8 | 54.9 | 61.9 | — | — |
| STEM & Reasoning 22 benchmarks | 83.1 | 71.9 | 64.1 | — | 49.9 | 67.6 | — | — | 60.3 | 35.9 | — | 52.7 | 3.8 | — | — |
| Reasoning 11 benchmarks | — | — | — | — | — | — | — | — | 0 | 80.9 | — | — | 70 | — | — |
| Document Understanding 9 benchmarks | 62.5 | — | 39.9 | 62.5 | — | 10.9 | 60.8 | 35.5 | — | 35 | 43.1 | 0 | 35 | — | — |
| General Capabilities 7 benchmarks | 70 | 0 | — | — | 53.8 | — | — | — | — | 0 | — | — | 85 | — | — |
| Video Understanding 7 benchmarks | — | — | 90.6 | 67.2 | — | — | — | — | — | — | 65 | — | — | — | — |
| Agentic 6 benchmarks | — | — | — | — | — | — | — | — | — | 25 | — | — | 75 | — | — |
| Knowledge 6 benchmarks | — | — | 85 | — | — | — | — | — | 45.4 | 30.5 | — | — | 55.6 | 54.9 | — |
| Spatial Intelligence 6 benchmarks | — | — | 100 | — | — | — | — | 68.2 | — | — | 39.2 | 21.5 | — | — | — |
| Finance 5 benchmarks | 73 | — | — | — | 48.4 | — | — | — | — | 27.3 | — | — | 54.7 | — | — |
| Instruction Following 4 benchmarks | — | 76.9 | — | — | — | — | — | — | — | 0 | — | 77.5 | 77.5 | — | — |
| Long Context 4 benchmarks | 44.3 | 77.5 | — | 77.5 | — | — | — | — | — | — | — | — | — | 26.8 | — |
| Code 3 benchmarks | — | — | — | — | — | — | — | — | — | 0 | — | — | 100 | — | — |
| Math 2 benchmarks | — | — | — | — | — | — | — | — | 100 | 49.6 | — | — | 72.8 | — | — |
| Multilingual 2 benchmarks | — | — | — | — | — | — | — | — | — | 0 | — | — | 100 | — | — |
| Classification 1 benchmark | — | — | — | — | — | — | — | — | — | 0 | — | — | 100 | — | — |
| Composite 1 benchmark | — | — | — | — | — | — | — | — | — | 0 | — | — | 100 | — | — |
| Cybersecurity 1 benchmark | — | — | — | — | — | — | — | — | — | 0 | — | — | 100 | — | — |
| Safety 1 benchmark | — | — | — | — | — | — | — | — | — | 0 | — | — | 100 | — | — |
Partial coverage benchmarks
Scored by at least one but not all selected models (224) — missing scores show as —
| Benchmark | Category | Claude-Sonnet-4.5 | Claude 4.5 Opus | Claude Opus 4.8 | Claude Opus 4.6 | Claude Sonnet 4.6 | Claude Opus 4.7 | DeepSeek V4 Pro | DeepSeek-V3.2 | Claude-Opus-4.5 | Claude-Opus-4.6 (Max) | Claude Opus 5 | Clef | Bonsai 2 27B — GGUF | Clef-Flash | d1-2B |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| SVG-Bench 3/15 models | coding_agent | — | — | — | — | 64.1 | 62.3 | 61.7 | — | — | — | — | — | — | — | — |
| Amazon ESCI 2/15 models | domain_finance | — | — | — | — | — | — | — | — | — | — | — | 57.5 | — | 57.4 | — |
| BPoMP 2/15 models | multilingual | — | — | — | — | — | — | — | — | — | — | — | 96.9 | — | 95.4 | — |
| HMMT Feb 25 1/15 models | stem_reasoning | — | 92.9 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| ForecastBench 2/15 models | reasoning | — | — | — | — | — | — | — | — | — | — | — | 13.9 | — | 10.6 | — |
| DynaMath 2/15 models | vision_language | 78.8 | 79.7 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| GPQA Diamond 8/15 models | stem_reasoning | — | 87 | 92 | 91.3 | — | — | 92.8 | — | — | 91.3 | 93.7 | 48 | — | 51 | — |
| IFBench (prompt loose) 1/15 models | instruction_following | — | — | — | — | — | — | — | — | — | — | — | — | 74 | — | — |
| MathVista (mini) 1/15 models | vision_language | 79.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| MMBench EN-DEV v1.1 2/15 models | vision_language | 88.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | 87.9 |
| ArXivMath 2/15 models | stem_reasoning | — | — | — | — | — | — | 62.1 | — | — | — | 71.5 | — | — | — | — |
| LiveCodeBench v6 4/15 models | stem_reasoning | — | 84.8 | — | 88.8 | — | — | — | — | — | 88.8 | — | — | 90.07 | — | — |
| WorkflowEvals 1/15 models | decision | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 81.6 |
| Hy-LifeSearch (Internal) 2/15 models | general_agent | — | — | — | — | — | — | 46.9 | — | — | — | 56.1 | — | — | — | — |
| FinEntity 2/15 models | domain_finance | — | — | — | — | — | — | — | — | — | — | — | 96.2 | — | 97.1 | — |
| ChessBench 2/15 models | reasoning | — | — | — | — | — | — | — | — | — | — | — | 24.7 | — | 23 | — |
| MRCR v2 256K (8-needle) 1/15 models | long_context | — | — | 83.2 | — | — | — | — | — | — | — | — | — | — | — | — |
| Fast Decisions 1/15 models | decision | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 58.8 |
| Automation-Bench 3/15 models | general_agent | — | — | 27.2 | — | — | — | 30.4 | — | — | — | 48.7 | — | — | — | — |
| SWE-bench Pro 10/15 models | coding_agent | 44 | 57.1 | 69.2 | 53.4 | — | 64.3 | 55.4 | 40.9 | 47 | 53.4 | 79.2 | — | — | — | — |
| NLI4CT 2/15 models | document_understanding | — | — | — | — | — | — | — | — | — | — | — | 82.9 | — | 78.6 | — |
| RealWorldQA 2/15 models | vision_language | 70.3 | — | — | — | — | — | — | — | — | 73.9 | — | — | — | — | — |
| BioMysteryBench 2/15 models | stem_reasoning | — | — | — | — | — | — | 61.6 | — | — | — | 72.1 | — | — | — | — |
| Home appliance simulator 2/15 models | general_agent | — | — | — | — | — | — | — | — | — | — | — | 83 | — | 97.7 | — |
| PRBench-Finance 1/15 models | general_capabilities | — | — | 51.9 | — | — | — | — | — | — | — | — | — | — | — | — |
| EmbSpatialBench 1/15 models | spatial_intelligence | 71.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| MathVision 3/15 models | vision_language | 71.1 | — | — | 71.2 | — | — | — | — | — | 65.5 | — | — | — | — | — |
| AIME 26 4/15 models | stem_reasoning | — | 95.1 | 95.7 | 96.7 | — | — | — | — | — | — | — | — | 95.83 | — | — |
| LiveSQLBench 1/15 models | coding_agent | — | — | — | — | — | 41 | — | — | — | — | — | — | — | — | — |
| AndroidWorld 1/15 models | general_agent | — | — | — | — | — | — | — | — | — | 62 | — | — | — | — | — |
| ACOS 2/15 models | general_capabilities | — | — | — | — | — | — | — | — | — | — | — | 33.3 | — | 25.9 | — |
| AgentWorldBench Terminal 3/15 models | general_agent | — | — | 59.18 | 57.51 | 56.98 | — | — | — | — | — | — | — | — | — | — |
| ContractNLI 2/15 models | document_understanding | — | — | — | — | — | — | — | — | — | — | — | 81.4 | — | 84.3 | — |
| Agents' Last Exam 3/15 models | general_agent | — | — | 27 | — | — | — | 21.9 | — | — | — | 25.1 | — | — | — | — |
| Toolathlon Verified 4/15 models | general_agent | — | — | 76.2 | 47.2 | — | — | 74.1 | — | — | — | 76.5 | — | — | — | — |
| CLINC150+OOS 2/15 models | classification | — | — | — | — | — | — | — | — | — | — | — | 97.4 | — | 66.8 | — |
| Hy-Backend 2.0 (Internal) 2/15 models | coding_agent | — | — | — | — | — | — | 33.8 | — | — | — | 40.3 | — | — | — | — |
| PLawBench 1/15 models | general_capabilities | — | — | 69.6 | — | — | — | — | — | — | — | — | — | — | — | — |
| ScreenSpot Pro 1/15 models | general_agent | 36.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| OJBench 1/15 models | stem_reasoning | — | — | — | 60.3 | — | — | — | — | — | — | — | — | — | — | — |
| Hy-SWE Max Verified (Internal) 2/15 models | coding_agent | — | — | — | — | — | — | 65.7 | — | — | — | 70.1 | — | — | — | — |
| AgentWorldBench Overall 3/15 models | general_agent | — | — | 56.59 | 57.8 | 56.04 | — | — | — | — | — | — | — | — | — | — |
| Hy-BrowseComp-Pro2 (Internal) 2/15 models | general_agent | — | — | — | — | — | — | 46.5 | — | — | — | 61.3 | — | — | — | — |
| ARC-c 2/15 models | reasoning | — | — | — | — | — | — | — | — | — | — | — | 97.7 | — | 98.3 | — |
| QwenReactBench 1/15 models | coding_agent | — | — | 1694 | — | — | — | — | — | — | — | — | — | — | — | — |
| MMStar 3/15 models | vision_language | 73.8 | 73.2 | — | — | — | — | — | — | — | — | — | — | — | — | 67.5 |
| GSM8K 3/15 models | math | — | — | — | — | — | — | — | — | — | — | — | 80.8 | 96.66 | 67.3 | — |
| ToolRet 2/15 models | agentic | — | — | — | — | — | — | — | — | — | — | — | 69.2 | — | 66.4 | — |
| Decision Bench (decisionbench.ai) 1/15 models | decision | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 85.2 |
| GDPVal 3/15 models | general_agent | — | — | — | — | 75.7 | 79.8 | 70.3 | — | — | — | — | — | — | — | — |
| OCRBench 1/15 models | document_understanding | 76.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| OfficeQA Pro 3/15 models | general_agent | — | — | — | — | — | 43.6 | 65.4 | — | — | — | 66.9 | — | — | — | — |
| PMC-VQA 1/15 models | vision_language | 55.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| QwenSVGBench 1/15 models | coding_agent | — | — | 1648 | — | — | — | — | — | — | — | — | — | — | — | — |
| DeepSearch QA 1/15 models | general_agent | — | — | — | 91.3 | — | — | — | — | — | — | — | — | — | — | — |
| RouterBench 2/15 models | general_agent | — | — | — | — | — | — | — | — | — | — | — | 79.7 | — | 79.9 | — |
| SWE Atlas - TW 4/15 models | coding_agent | — | — | — | — | 31.8 | 38.2 | 45.6 | — | — | — | 68.7 | — | — | — | — |
| CL-bench 1/15 models | coding_agent | — | — | — | — | — | 22.9 | — | — | — | — | — | — | — | — | — |
| MBPP+ 1/15 models | code | — | — | — | — | — | — | — | — | — | — | — | — | 83.07 | — | — |
| LOCA-Bench (256k) 2/15 models | long_context | — | — | — | — | — | 57 | — | — | — | — | 53.7 | — | — | — | — |
| KernelBench Hard 1/15 models | coding_agent | — | — | — | — | — | 30.7 | — | — | — | — | — | — | — | — | — |
| AgentWorldBench Web 3/15 models | general_agent | — | — | 54.66 | 51.42 | 50.78 | — | — | — | — | — | — | — | — | — | — |
| SWE Atlas - QnA 4/15 models | coding_agent | — | — | — | — | 31.2 | 45.2 | 53.4 | — | — | — | 57.5 | — | — | — | — |
| AI2D_TEST 2/15 models | document_understanding | 87 | — | — | — | — | — | — | — | — | — | — | — | — | — | 76.7 |
| LingoQA 1/15 models | spatial_intelligence | 12.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| IMO 2025 1/15 models | stem_reasoning | — | — | — | — | — | 17.4 | — | — | — | — | — | — | — | — | — |
| NL2Repo 4/15 models | coding_agent | — | 43.2 | 69.7 | — | — | — | 61.5 | — | — | — | 75.3 | — | — | — | — |
| PRBench-Legal 1/15 models | general_capabilities | — | — | 52.7 | — | — | — | — | — | — | — | — | — | — | — | — |
| MMMU-Pro 7/15 models | vision_language | 68.4 | 70.6 | — | 73.9 | 74.5 | 77 | — | — | — | — | — | — | 75.49 | — | 26.4 |
| Program Bench 1/15 models | coding_agent | — | — | 63.8 | — | — | — | — | — | — | — | — | — | — | — | — |
| MCPMark-Verified 1/15 models | agentic | — | — | 76.4 | — | — | — | — | — | — | — | — | — | — | — | — |
| PostTrainBench 4/15 models | coding_agent | — | — | 37.2 | — | — | 42.4 | 24.5 | — | — | — | 35 | — | — | — | — |
| SGD/SGD-X 2/15 models | instruction_following | — | — | — | — | — | — | — | — | — | — | — | 43.8 | — | 34.2 | — |
| QwenClawBench 1/15 models | coding_agent | — | 52.3 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| BabyVision 3/15 models | vision_language | 18.6 | — | — | 14.8 | — | — | — | — | — | — | 94.1 | — | — | — | — |
| SLAKE 1/15 models | vision_language | 73.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| MMMLU 1/15 models | multilingual | — | — | — | — | — | — | — | 87.9 | — | — | — | — | — | — | — |
| Claw-Eval Avg 6/15 models | coding_agent | — | 76.6 | — | 82.4 | 68.3 | 71.6 | 58.4 | — | — | 54.7 | — | — | — | — | — |
| ANLI 2/15 models | reasoning | — | — | — | — | — | — | — | — | — | — | — | 69.8 | — | 59.1 | — |
| VideoMMMU 3/15 models | video_understanding | 77.6 | 84.4 | — | — | — | 83 | — | — | — | — | — | — | — | — | — |
| HMMT Nov 25 2/15 models | stem_reasoning | — | 93.3 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — |
| Apex-Agents 5/15 models | general_agent | — | — | — | 33 | 26.2 | 37.2 | 32.4 | — | — | — | 41.8 | — | — | — | — |
| HumanEval+ 1/15 models | code | — | — | — | — | — | — | — | — | — | — | — | — | 95.12 | — | — |
| ARC-e 2/15 models | reasoning | — | — | — | — | — | — | — | — | — | — | — | 99 | — | 99.5 | — |
| VisualWebBench 1/15 models | vision_language | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 96.2 |
| MMRB2 1/15 models | multimodal | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 54.2 |
| WorkSpaceBench 3/15 models | general_agent | — | — | 66.8 | — | — | — | 65.4 | — | — | — | 75 | — | — | — | — |
| MMLU-Pro 4/15 models | knowledge | — | 89.5 | — | — | — | — | — | 65.5 | — | — | — | 65.9 | — | 65.3 | — |
| NaturalBench 1/15 models | multimodal | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 81.5 |
| BankerToolBench 3/15 models | general_agent | — | — | — | — | — | 81.3 | 73.1 | — | — | — | 81.9 | — | — | — | — |
| SuperGPQA 2/15 models | knowledge | — | 70.6 | — | — | — | — | — | 45 | — | — | — | — | — | — | — |
| E-Bench-Code (Internal) 2/15 models | coding_agent | — | — | — | — | — | — | 64.3 | — | — | — | 82.7 | — | — | — | — |
| OSWorld-Verified 4/15 models | general_agent | 61.4 | — | — | 72.7 | 72.5 | 82.8 | — | — | — | — | — | — | — | — | — |
| TIR-Bench 1/15 models | vision_language | 27.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| AgentWorldBench Search 3/15 models | general_agent | — | — | 35.14 | 29.3 | 28.79 | — | — | — | — | — | — | — | — | — | — |
| POP909-CL 2/15 models | general_capabilities | — | — | — | — | — | — | — | — | — | — | — | 15.8 | — | 1.6 | — |
| CV-Bench 1/15 models | vision_language | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 86.1 |
| HellaSwag 2/15 models | reasoning | — | — | — | — | — | — | — | — | — | — | — | 98.2 | — | 98.6 | — |
| IFEval 1/15 models | instruction_following | — | — | — | — | — | — | — | — | — | — | — | — | 91.31 | — | — |
| HLE (with tools) 5/15 models | stem_reasoning | — | — | 57.9 | 53 | — | — | 60 | — | — | 40 | 60.9 | — | — | — | — |
| LVBench 1/15 models | video_understanding | — | — | — | — | — | — | — | — | — | 63 | — | — | — | — | — |
| CharXiv (RQ) 5/15 models | document_understanding | 67.2 | 68.5 | — | 69.1 | — | — | — | — | — | 66 | 83.7 | — | — | — | — |
| Kimi Claw 24/7 Bench 1/15 models | agentic | — | — | 50.4 | — | — | — | — | — | — | — | — | — | — | — | — |
| POPE 1/15 models | vision_language | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 88.7 |
| CLadder 2/15 models | stem_reasoning | — | — | — | — | — | — | — | — | — | — | — | 94 | — | 97.7 | — |
| HoVer 2/15 models | knowledge | — | — | — | — | — | — | — | — | — | — | — | 65.2 | — | 61.2 | — |
| FrontierSWE 1/15 models | coding_agent | — | — | 70 | — | — | — | — | — | — | — | — | — | — | — | — |
| GDP.pdf 1/15 models | long_context | — | — | — | — | — | — | — | — | — | — | 21.6 | — | — | — | — |
| cfcolor 2/15 models | vision_language | — | — | — | — | — | — | — | — | — | — | — | 66 | — | 65.8 | — |
| CritPt (no tools) 3/15 models | stem_reasoning | — | — | 20.9 | — | — | — | 18 | — | — | — | 29.1 | — | — | — | — |
| BrowseComp 4/15 models | general_agent | — | — | — | 83.7 | 74.7 | 79.3 | 83.4 | — | — | — | — | — | — | — | — |
| MedXpertQA-MM 1/15 models | vision_language | 54 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| MMLU-Redux 3/15 models | knowledge | — | 95.6 | — | — | — | — | — | 87.5 | — | — | — | — | 89.09 | — | — |
| SATA-Bench 2/15 models | general_agent | — | — | — | — | — | — | — | — | — | — | — | 33.8 | — | 36.7 | — |
| NL2Repo-Bench 5/15 models | coding_agent | — | — | 69.4 | — | — | 56.3 | 35.5 | — | — | 47.6 | 75.3 | — | — | — | — |
| BRIGHT 2/15 models | agentic | — | — | — | — | — | — | — | — | — | — | — | 45.9 | — | 39.3 | — |
| DRACO 4/15 models | general_agent | — | — | — | — | 75.8 | 77.7 | 77.3 | — | — | — | 88.6 | — | — | — | — |
| SkillsBench 1/15 models | general_agent | — | — | 65.1 | — | — | — | — | — | — | — | — | — | — | — | — |
| CoWorkBench 2/15 models | general_agent | — | — | 72.3 | — | — | — | — | — | — | 68.2 | — | — | — | — | — |
| ImajevBench 1/15 models | multimodal | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 75.5 |
| MCPMark 1/15 models | general_agent | — | — | — | 56.7 | — | — | — | — | — | — | — | — | — | — | — |
| SWE-bench Multilingual 6/15 models | coding_agent | — | 77.5 | — | 77.8 | — | — | 77.3 | 62.3 | — | 77.5 | 89.5 | — | — | — | — |
| MMLU 3/15 models | knowledge | — | — | — | — | — | — | — | 87.8 | — | — | — | 90.3 | — | 91.8 | — |
| MME 1/15 models | vision_language | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 85.4 |
| VAST 2/15 models | vision_language | — | — | — | — | — | — | — | — | — | — | — | 59.5 | — | 49.6 | — |
| MMVU 1/15 models | video_understanding | 70.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| PhishNChips 2/15 models | cybersecurity | — | — | — | — | — | — | — | — | — | — | — | 79.6 | — | 75 | — |
| Decision Index (public suite) 1/15 models | decision | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 51.1 |
| PaperBench 2/15 models | coding_agent | — | — | 80.3 | — | — | 58.5 | — | — | — | — | — | — | — | — | — |
| YC-Bench 3/15 models | general_agent | — | — | — | — | 100000 | 2200000 | 500000 | — | — | — | — | — | — | — | — |
| MLVU 2/15 models | video_understanding | 72.8 | 81.7 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| OCR Bench v2 1/15 models | document_understanding | — | — | — | — | — | — | — | — | — | — | — | — | 56.88 | — | — |
| JevBench 1/15 models | decision | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 77.9 |
| API-Bank 2/15 models | agentic | — | — | — | — | — | — | — | — | — | — | — | 91.9 | — | 93.1 | — |
| Hy-FinAgentBench (Internal) 2/15 models | domain_finance | — | — | — | — | — | — | 78.5 | — | — | — | 82 | — | — | — | — |
| IFBench 2/15 models | instruction_following | — | — | 62.2 | — | — | — | — | — | — | 62.5 | — | — | — | — | — |
| SWE-bench Verified 6/15 models | coding_agent | — | 80.9 | — | 80.8 | 79.6 | 87.6 | 80.6 | 70.2 | — | — | — | — | — | — | — |
| IMOAnswerBench 3/15 models | stem_reasoning | — | 84 | 83.5 | 75.3 | — | — | — | — | — | — | — | — | — | — | — |
| Humanity's Last Exam 5/15 models | stem_reasoning | — | 30.8 | 45.7 | 40 | — | — | 42.7 | — | — | — | 54.9 | — | — | — | — |
| CountBench 2/15 models | spatial_intelligence | 90 | 90.6 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| VL-RewardBench 1/15 models | vision_language | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 69.3 |
| CountQA 1/15 models | spatial_intelligence | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 51.4 |
| ERQA 4/15 models | spatial_intelligence | 45 | 46.8 | — | — | — | — | — | — | — | 40.8 | — | — | — | — | 45.8 |
| VideoMME (w/o sub.) 1/15 models | video_understanding | 75.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| Terminal-Bench 2.0 3/15 models | coding_agent | — | 59.3 | — | 65.4 | — | — | — | 39.3 | — | — | — | — | — | — | — |
| VisualPuzzles 1/15 models | vision_language | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 29.9 |
| CRUXEval 2/15 models | code | — | — | — | — | — | — | — | — | — | — | — | 86.7 | — | 86.1 | — |
| BIG-Bench Hard 2/15 models | reasoning | — | — | — | — | — | — | — | — | — | — | — | 73.7 | — | 68.9 | — |
| QwenSWEBench 1/15 models | coding_agent | — | — | 84 | — | — | — | — | — | — | — | — | — | — | — | — |
| SpreadSheetBench-v1 2/15 models | general_agent | — | — | — | — | — | 88.5 | 84.9 | — | — | — | — | — | — | — | — |
| QwenWebBench 1/15 models | coding_agent | — | 1536 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| VlmsAreBlind 1/15 models | vision_language | 85.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| WinoGrande 2/15 models | reasoning | — | — | — | — | — | — | — | — | — | — | — | 93.5 | — | 97.5 | — |
| When2Call MCQ 2/15 models | agentic | — | — | — | — | — | — | — | — | — | — | — | 72.4 | — | 65.6 | — |
| SUPERChem 2/15 models | stem_reasoning | — | — | — | — | — | — | 62 | — | — | — | 76.7 | — | — | — | — |
| VIBE-V2 3/15 models | coding_agent | — | — | — | — | 42.8 | 55.9 | 45 | — | — | — | — | — | — | — | — |
| MATH-500 1/15 models | math | — | — | — | — | — | — | — | — | — | — | — | — | 98.8 | — | — |
| HorizonMath (pass@4) 2/15 models | stem_reasoning | — | — | — | — | — | — | 4.42 | — | — | — | 5.3 | — | — | — | — |
| Terminal-Bench 2.1 (Terminus-2) 3/15 models | coding_agent | — | — | 85 | — | — | — | 87.9 | — | — | — | 86.7 | — | — | — | — |
| RAGTruth 2/15 models | safety | — | — | — | — | — | — | — | — | — | — | — | 79.4 | — | 35.6 | — |
| DeepSWE 2/15 models | coding_agent | — | — | — | — | — | — | 62.7 | — | — | — | 68.8 | — | — | — | — |
| BrokenArXiv 2/15 models | stem_reasoning | — | — | — | — | — | — | 43.1 | — | — | — | 77.7 | — | — | — | — |
| OmniDocBench 1.5 3/15 models | document_understanding | 85.8 | — | — | — | 86.9 | 89.3 | — | — | — | — | — | — | — | — | — |
| HallusionBench 2/15 models | vision_language | 59.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | 72.1 |
| OCRBench 1/15 models | document_understanding | 76.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| HMMT Feb 26 3/15 models | stem_reasoning | — | 85.3 | 96.7 | 96.2 | — | — | — | — | — | — | — | — | — | — | — |
| BANKING77 2/15 models | domain_finance | — | — | — | — | — | — | — | — | — | — | — | 94.2 | — | 90.9 | — |
| AgentWorldBench OS 3/15 models | general_agent | — | — | 66.62 | 70.2 | 63.17 | — | — | — | — | — | — | — | — | — | — |
| BLINK 1/15 models | multimodal | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 66.2 |
| MMVP 1/15 models | vision_language | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 82.7 |
| AgentWorldBench MCP 3/15 models | general_agent | — | — | 54.93 | 69.9 | 70 | — | — | — | — | — | — | — | — | — | — |
| Hy-FinmodelBench v2 (Internal) 2/15 models | domain_finance | — | — | — | — | — | — | 51.3 | — | — | — | 66 | — | — | — | — |
| MathArena Apex 2025 2/15 models | stem_reasoning | — | — | — | — | — | — | 66.3 | — | — | — | 91.4 | — | — | — | — |
| Terminal Bench 2.1 3/15 models | coding_agent | — | — | 84.6 | — | — | 66.1 | 59.6 | — | — | — | — | — | — | — | — |
| AIME 2025 1/15 models | stem_reasoning | — | — | — | — | — | — | — | — | — | — | — | — | 95 | — | — |
| AgentWorldBench SWE 3/15 models | general_agent | — | — | 64.1 | 64.55 | 64.52 | — | — | — | — | — | — | — | — | — | — |
| CC-OCR 2/15 models | document_understanding | 68.1 | 76.9 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| JobBench 4/15 models | general_agent | — | — | 48.4 | — | — | — | 54.1 | — | — | 36.6 | 68 | — | — | — | — |
| $OneMillion-Bench 3/15 models | general_capabilities | — | — | 41.8 | — | — | — | 62 | — | — | — | 68.1 | — | — | — | — |
| Kimi Code Bench V2 1/15 models | coding_agent | — | — | 67.4 | — | — | — | — | — | — | — | — | — | — | — | — |
| SciCode (subtask) 1/15 models | stem_reasoning | — | — | — | 51.9 | — | — | — | — | — | — | — | — | — | — | — |
| USAMO 2026 1/15 models | stem_reasoning | — | — | — | — | — | 52.8 | — | — | — | — | — | — | — | — | — |
| Tool Decathlon 1/15 models | general_agent | — | — | 59.9 | — | — | — | — | — | — | — | — | — | — | — | — |
| QwenQoderBench 1/15 models | coding_agent | — | — | 62.7 | — | — | — | — | — | — | — | — | — | — | — | — |
| MuSR 3/15 models | reasoning | — | — | — | — | — | — | — | — | — | — | — | 83.5 | 70.63 | 86 | — |
| OSWorld 2.0 (Binary) 1/15 models | general_agent | — | — | — | — | — | — | — | — | — | — | 70.6 | — | — | — | — |
| Aider 1/15 models | coding_agent | — | — | — | — | — | — | — | 69.9 | — | — | — | — | — | — | — |
| ZEROBench_sub 1/15 models | vision_language | 26.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| C-Eval 2/15 models | knowledge | — | 92.2 | — | — | — | — | — | 90.4 | — | — | — | — | — | — | — |
| Humicroedit 2/15 models | reasoning | — | — | — | — | — | — | — | — | — | — | — | 66.7 | — | 75.1 | — |
| WideSearch 3/15 models | general_agent | — | — | 72.9 | — | — | — | 81.8 | — | — | — | 84 | — | — | — | — |
| SimpleVQA 2/15 models | vision_language | 57.6 | 65.7 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| VideoMME (w sub.) 2/15 models | video_understanding | 81.1 | 77.7 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| SkillsBench Avg5 3/15 models | coding_agent | — | 45.3 | — | — | — | — | 65 | — | — | — | 63.7 | — | — | — | — |
| New Yorker 2/15 models | composite | — | — | — | — | — | — | — | — | — | — | — | 69.5 | — | 66.1 | — |
| Habermas Machine 2/15 models | reasoning | — | — | — | — | — | — | — | — | — | — | — | 68.7 | — | 71.8 | — |
| HealthBench 1/15 models | general_capabilities | — | — | 52.4 | — | — | — | — | — | — | — | — | — | — | — | — |
| Claw-Eval Pass^3 3/15 models | coding_agent | — | 59.6 | — | 70.4 | — | — | — | — | — | 52.5 | — | — | — | — | — |
| E-Bench (Internal) 2/15 models | general_agent | — | — | — | — | — | — | 61.3 | — | — | — | 77.8 | — | — | — | — |
| ProgramBench 3/15 models | coding_agent | — | — | 71.9 | — | — | — | 15.5 | — | — | — | 39.5 | — | — | — | — |
| LongBench v2 2/15 models | long_context | — | — | 69.1 | — | — | — | — | 40.2 | — | — | — | — | — | — | — |
| MCP-Atlas 5/15 models | general_agent | — | — | 77.8 | — | 61.3 | 77 | 73.6 | — | — | — | 85.7 | — | — | — | — |
| AndroidBench 1/15 models | coding_agent | — | — | 69.8 | — | — | — | — | — | — | — | — | — | — | — | — |
| SWE-Marathon 3/15 models | coding_agent | — | — | 26 | — | — | — | 19 | — | — | — | 50 | — | — | — | — |
| Terminal-Bench 2.1 (Best Reported Harness) 1/15 models | coding_agent | — | — | 78.9 | — | — | — | — | — | — | — | — | — | — | — | — |
| Cybergym 1/15 models | general_agent | — | — | — | — | — | — | 83.3 | — | — | — | — | — | — | — | — |
| Terminal-Bench 4.0 1/15 models | coding_agent | — | — | — | — | — | — | — | — | — | — | 51.8 | — | — | — | — |
| DecisionBench (Hanno Labs) 1/15 models | decision | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 77.1 |
| SWE Atlas - RF 2/15 models | coding_agent | — | — | — | — | — | — | 48.6 | — | — | — | 60 | — | — | — | — |
| V-Star 4/15 models | vision_language | 58.6 | 67 | — | 86.4 | — | — | — | — | — | — | — | — | — | — | 76.4 |
| AgentWorldBench Android 3/15 models | general_agent | — | — | 61.5 | 61.74 | 58.03 | — | — | — | — | — | — | — | — | — | — |
| DeepSWE 1.1 2/15 models | coding_agent | — | — | 59 | — | — | — | — | — | — | — | 74 | — | — | — | — |
| Typed Decisions 1/15 models | decision | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 61.3 |
| Hy-CompanyBench V2 (Internal) 2/15 models | general_agent | — | — | — | — | — | — | 64.4 | — | — | — | 72.7 | — | — | — | — |
| Harbor-Index 2/15 models | coding_agent | — | — | — | — | — | — | 36.9 | — | — | — | 56.9 | — | — | — | — |
| SWE-efficiency 1/15 models | coding_agent | — | — | — | — | — | 42.2 | — | — | — | — | — | — | — | — | — |
| SWE Refactor Bench 1/15 models | coding_agent | — | — | — | — | — | — | — | — | — | — | 31 | — | — | — | — |
| MLS-Bench-Lite 1/15 models | coding_agent | — | — | 42.8 | — | — | — | — | — | — | — | — | — | — | — | — |
| BFCL-V3 1/15 models | general_agent | — | — | — | — | — | — | — | — | — | — | — | — | 74.92 | — | — |
| RefSpatialBench 1/15 models | spatial_intelligence | 2.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| MMMU 2/15 models | vision_language | 79.6 | 80.7 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| GDPVal-AA v2 3/15 models | general_agent | — | — | 1582 | — | — | — | 1580 | — | — | — | 1831 | — | — | — | — |
| MVBench 1/15 models | video_understanding | — | 67.2 | — | — | — | — | — | — | — | — | — | — | — | — | — |
| ZEROBench 1/15 models | vision_language | 4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
Comparison FAQ
Which model scores highest in this comparison?
Claude Opus 5 scores higher overall across category strength profiles (60 vs 52 for Claude Opus 4.8); only 0 benchmarks are directly comparable.
Which model is strongest at General Agents?
Claude Opus 5 leads the General Agents category (strength score 76.3/100 across 33 comparable benchmarks).
How reliable is this comparison verdict?
Only 0 of 224 benchmarks are scored by every selected model — the verdict leans on category strength profiles and is indicative.