Agentic · tools
CyberGym leaderboard
As of 2026-11-21, the highest measured score on CyberGym is 95.1% by MiMo-V2.6-Flash. 180 more models have estimated scores, calibrated from the benchmarks they were measured on.
Measured scores: benchlm.ai.
| # | Model | Score | Source |
|---|---|---|---|
| 1 | Muse Wonder Full | 97.7% | estimated ± 6.0 pp, low confidence |
| 2 | Kimi K3.1 Pro | 96.3% | estimated ± 5.9 pp, low confidence |
| 3 | MiMo-V2.6-Flash | 95.1% | measured |
| 4 | Claude Opus 5 | 94.8% | estimated ± 4.5 pp, medium confidence |
| 5 | Muse Wonder | 94.4% | estimated ± 6.0 pp, low confidence |
| 6 | MiMo-V2.6-Pro | 94.0% | measured |
| 7 | Claude Sonnet 5.5 | 92.7% | estimated ± 6.0 pp, low confidence |
| 8 | ERNIE 5.5 Flash Preview | 89.3% | estimated ± 4.6 pp, medium confidence |
| 9 | Hy4 | 88.3% | estimated ± 4.6 pp, medium confidence |
| 10 | Claude Haiku 5.5 | 88.2% | estimated ± 6.0 pp, medium confidence |
| 11 | GPT-5.5 Pro | 88.2% | estimated ± 7.3 pp, medium confidence |
| 12 | DeepSeek V4.1 Flash 1017 | 88.1% | measured |
| 13 | Gemini 4 Carbon | 87.9% | estimated ± 4.6 pp, medium confidence |
| 14 | Ling 3.1 Flash | 87.9% | measured |
| 15 | Hy4 flash | 87.6% | estimated ± 4.6 pp, medium confidence |
| 16 | LongCat-2.5 | 87.5% | estimated ± 4.6 pp, medium confidence |
| 17 | Claude Fable 5.2 | 87.4% | estimated ± 4.6 pp, medium confidence |
| 18 | Gemini 4 Argon | 87.4% | estimated ± 4.6 pp, medium confidence |
| 19 | Ultramarine 1 Flash | 87.1% | estimated ± 4.6 pp, medium confidence |
| 20 | Fugu Cyber | 86.9% | measured |
| 21 | GPT-5.4 Pro | 86.7% | estimated ± 7.3 pp, medium confidence |
| 22 | GPT-6 Cyber | 86.6% | estimated ± 10.0 pp, low confidence |
| 23 | Atria Dawn Preview | 86.5% | measured |
| 24 | Claude Mythos 5.2 | 86.5% | estimated ± 10.0 pp, low confidence |
| 25 | GPT-5.6 Cyber | 86.3% | estimated ± 10.0 pp, medium confidence |
| 26 | Gemini 3.8 Flash Cyber | 86.2% | measured |
| 27 | Claude Mythos 5.1 | 86.0% | estimated ± 10.0 pp, medium confidence |
| 28 | Gemini 3.8 Flash | 85.6% | estimated ± 4.5 pp, high confidence |
| 29 | Grok 4.6 | 85.4% | estimated ± 3.1 pp, low confidence |
| 30 | Muse Spark 1.3 | 85.4% | estimated ± 3.1 pp, low confidence |
| 31 | GPT-6 Astra | 84.9% | estimated ± 4.6 pp, medium confidence |
| 32 | LFM3-1.7B | 84.8% | estimated ± 4.6 pp, medium confidence |
| 33 | Step 5 Preview | 84.7% | measured |
| 34 | GLM-5.3 | 84.5% | measured |
| 35 | GPT-5.6 Sol | 84.5% | measured |
| 36 | Qwen3.8-27B | 84.4% | estimated ± 3.1 pp, low confidence |
| 37 | Claude Opus 5.5 | 84.4% | estimated ± 4.6 pp, medium confidence |
| 38 | Holo3-35B-A3B | 84.3% | estimated ± 11.3 pp, low confidence |
| 39 | Claude Fable 5.1 | 84.1% | estimated ± 3.1 pp, low confidence |
| 40 | GLM-5.3-Flash | 84.1% | estimated ± 3.1 pp, low confidence |
| 41 | Claude Mythos 5 | 83.8% | measured |
| 42 | Qwen3.8 Max Preview | 83.6% | estimated ± 5.3 pp, medium confidence |
| 43 | Kimi K3 | 83.6% | estimated ± 3.1 pp, low confidence |
| 44 | Muse Spark Open | 83.5% | estimated ± 6.0 pp, low confidence |
| 45 | Qwen3.8-Flash-Next | 83.4% | estimated ± 3.1 pp, low confidence |
| 46 | DeepSeek V4 Pro | 83.3% | measured |
| 47 | Gemini 3.5 Flash Cyber | 83.2% | measured |
| 48 | Ripple-65B Preview | 83.1% | estimated ± 4.6 pp, medium confidence |
| 49 | Claude Mythos Preview | 83.1% | measured |
| 50 | GPT-6.1 Sol | 83.0% | estimated ± 4.6 pp, medium confidence |
| 51 | MiniMax M3.1 Flash | 82.2% | measured |
| 52 | Grok 4.5 | 81.9% | estimated ± 3.1 pp, low confidence |
| 53 | GPT-6 Sol | 81.8% | estimated ± 4.6 pp, medium confidence |
| 54 | GPT-5.5 | 81.8% | measured |
| 55 | GPT-5.6 Terra | 81.8% | measured |
| 56 | Inkling 2 Small | 81.1% | estimated ± 3.1 pp, low confidence |
| 57 | Ling 3.0 Flash Fin | 80.2% | estimated ± 3.1 pp, low confidence |
| 58 | Ornith-1.5-397B | 80.0% | estimated ± 6.0 pp, low confidence |
| 59 | Claude Fable 5 | 79.9% | estimated ± 3.1 pp, low confidence |
| 60 | Gemini 3.7 Flash | 79.5% | estimated ± 4.5 pp, high confidence |
| 61 | Claude Sonnet 5 | 79.5% | estimated ± 3.1 pp, low confidence |
| 62 | GPT-6 Luna | 79.4% | estimated ± 6.0 pp, medium confidence |
| 63 | Muse Spark 1.2 | 79.3% | estimated ± 5.3 pp, medium confidence |
| 64 | GPT-5.4 | 79.0% | measured |
| 65 | LFM2.5-24B-A2B | 79.0% | estimated ± 4.6 pp, medium confidence |
| 66 | Mistral Large 4 | 78.9% | estimated ± 6.0 pp, medium confidence |
| 67 | Holo3-122B-A10B | 78.8% | estimated ± 11.3 pp, low confidence |
| 68 | Qwen3.8 Max | 78.7% | estimated ± 4.6 pp, medium confidence |
| 69 | Hy4 preview | 78.4% | measured |
| 70 | Claude Opus 4.8 | 78.3% | estimated ± 5.3 pp, medium confidence |
| 71 | GPT-5.6 Luna | 77.9% | measured |
| 72 | Grok 4.7 | 76.9% | estimated ± 5.8 pp, low confidence |
| 73 | dots3-note Preview | 76.7% | estimated ± 7.3 pp, medium confidence |
| 74 | DeepSeek V4 Flash 0731 | 76.7% | measured |
| 75 | GLM-5.2 | 75.8% | estimated ± 5.3 pp, medium confidence |
| 76 | UI-Mate-27B | 75.6% | estimated ± 11.3 pp, low confidence |
| 77 | Sage-A1.5 | 74.3% | estimated ± 7.3 pp, medium confidence |
| 78 | Solar Open 2 | 73.6% | estimated ± 9.8 pp, medium confidence |
| 79 | Claude Opus 4.7 | 73.1% | measured |
| 80 | Ling 3.0 Flash | 73.1% | estimated ± 3.1 pp, low confidence |
| 81 | Ling 3.0 Flash Sante | 72.0% | estimated ± 7.3 pp, medium confidence |
| 82 | Qwen3.5-27B | 71.3% | estimated ± 5.8 pp, low confidence |
| 83 | Quasar 438B | 70.7% | estimated ± 5.3 pp, medium confidence |
| 84 | Inkling 2 Mini | 70.2% | estimated ± 7.3 pp, medium confidence |
| 85 | MiniMax M3 | 69.2% | estimated ± 5.3 pp, medium confidence |
| 86 | Gemini 3.6 Flash | 68.7% | estimated ± 5.3 pp, medium confidence |
| 87 | GLM-5.1 | 68.7% | measured |
| 88 | Beam 1.0 501B-A23B FP8 | 66.9% | estimated ± 7.3 pp, medium confidence |
| 89 | Claude Opus 4.6 | 66.6% | measured |
| 90 | Gemini 3.5 Flash | 66.5% | estimated ± 5.3 pp, medium confidence |
| 91 | Gemini 3 Pro | 66.1% | estimated ± 9.3 pp, medium confidence |
| 92 | Apodex 1.1 | 65.6% | estimated ± 6.0 pp, medium confidence |
| 93 | Apodex 1.1 Mini | 65.6% | estimated ± 6.0 pp, medium confidence |
| 94 | Beam 1.0 501B-A23B NVFP4 | 65.3% | estimated ± 7.3 pp, medium confidence |
| 95 | Hy3 | 65.3% | estimated ± 5.3 pp, medium confidence |
| 96 | Hy3 Preview | 65.3% | estimated ± 5.3 pp, medium confidence |
| 97 | Agents-A1 | 65.2% | estimated ± 7.3 pp, medium confidence |
| 98 | Claude Sonnet 4.6 | 65.2% | measured |
| 99 | Beam 1.0 501B-A23B | 64.9% | estimated ± 6.0 pp, medium confidence |
| 100 | Inkling-Small | 64.7% | estimated ± 5.3 pp, low confidence |
| 101 | Inkling | 64.2% | estimated ± 5.3 pp, low confidence |
| 102 | Qwen3.7 Max | 63.9% | estimated ± 5.3 pp, low confidence |
| 103 | Nemotron 3.5 Super VL 120B A12B | 63.9% | estimated ± 6.0 pp, medium confidence |
| 104 | Nemotron 3.5 Super VL 120B A12B NVFP4 | 63.9% | estimated ± 6.0 pp, medium confidence |
| 105 | Ling 3.0 Flash VL | 63.8% | estimated ± 6.0 pp, medium confidence |
| 106 | MiMo-V2.5-Pro | 63.0% | estimated ± 5.3 pp, low confidence |
| 107 | Kimi K2.7 Code | 62.8% | estimated ± 5.3 pp, low confidence |
| 108 | Kimi K2.6 | 62.6% | estimated ± 5.3 pp, low confidence |
| 109 | Nemotron 3 Ultra | 62.2% | estimated ± 5.3 pp, low confidence |
| 110 | Qwen3.6-27B | 61.0% | estimated ± 5.3 pp, low confidence |
| 111 | Qwen3.7 Plus | 60.7% | estimated ± 5.3 pp, low confidence |
| 112 | GPT-5.4 mini | 60.6% | estimated ± 5.3 pp, low confidence |
| 113 | Solar Pro 4 | 60.2% | estimated ± 6.0 pp, medium confidence |
| 114 | GPT-5.4 nano | 59.2% | estimated ± 5.3 pp, low confidence |
| 115 | Muse Spark 1.1 | 59.0% | measured |
| 116 | Grok 4.3 | 58.8% | estimated ± 5.3 pp, low confidence |
| 117 | MiniMax M2.7 | 58.5% | estimated ± 5.3 pp, low confidence |
| 118 | Nemotron 3.5 Bolt 74B-A9B BF16 | 57.9% | estimated ± 7.3 pp, medium confidence |
| 119 | Gemini 3.5 Flash-Lite | 57.8% | estimated ± 5.3 pp, low confidence |
| 120 | Qwen3.6-35B-A3B | 57.1% | estimated ± 5.3 pp, low confidence |
| 121 | LLaDA2.2-flash | 56.7% | estimated ± 9.8 pp, medium confidence |
| 122 | GPT-5.3 Codex | 56.7% | estimated ± 9.3 pp, medium confidence |
| 123 | Qwen3.5 397B | 56.6% | estimated ± 7.3 pp, medium confidence |
| 124 | Gemini 3 Flash | 55.3% | estimated ± 9.3 pp, medium confidence |
| 125 | Ornith-1.5-35B-A3B | 54.9% | estimated ± 7.3 pp, medium confidence |
| 126 | Agents-A1-4B | 53.9% | estimated ± 7.3 pp, medium confidence |
| 127 | Step 3.7 Flash | 53.7% | estimated ± 6.0 pp, medium confidence |
| 128 | Muse Glimmer 30B | 53.6% | estimated ± 5.3 pp, low confidence |
| 129 | GLM-4.7 | 53.5% | estimated ± 6.0 pp, medium confidence |
| 130 | Gemini 3.1 Pro | 53.5% | estimated ± 5.3 pp, low confidence |
| 131 | Qwen3.5-122B-A10B | 53.0% | estimated ± 5.3 pp, low confidence |
| 132 | Mistral Medium 3.5 128B | 52.7% | estimated ± 5.3 pp, low confidence |
| 133 | GPT-5.2 | 52.7% | estimated ± 7.3 pp, medium confidence |
| 134 | UI-Mate-9B | 52.4% | estimated ± 11.3 pp, low confidence |
| 135 | Qwen3.6 Plus | 51.7% | estimated ± 6.0 pp, low confidence |
| 136 | Nemotron 3.5 Bolt 74B-A9B | 50.9% | estimated ± 6.0 pp, low confidence |
| 137 | Gemma 4 31B | 50.7% | estimated ± 5.3 pp, low confidence |
| 138 | Claude Opus 4.5 | 50.6% | measured |
| 139 | GPT-OSS 120B | 50.3% | estimated ± 5.3 pp, low confidence |
| 140 | Nemotron 3.5 Lightning 30B A3B NVFP4 | 50.3% | estimated ± 5.3 pp, low confidence |
| 141 | A.X K2 | 49.3% | estimated ± 6.0 pp, low confidence |
| 142 | Granite 4.2 8B | 48.4% | estimated ± 5.3 pp, low confidence |
| 143 | Command A+ | 48.3% | estimated ± 5.3 pp, low confidence |
| 144 | Gemini 2.5 Pro | 48.3% | estimated ± 5.3 pp, low confidence |
| 145 | GPT-5.2-Codex | 48.0% | estimated ± 9.3 pp, medium confidence |
| 146 | Claude 4.1 Opus | 47.6% | estimated ± 11.4 pp, low confidence |
| 147 | Mistral Large 3 | 47.4% | estimated ± 5.3 pp, low confidence |
| 148 | Qwen3.5-35B-A3B | 47.2% | estimated ± 7.3 pp, low confidence |
| 149 | Mistral Small 4 | 46.7% | estimated ± 5.3 pp, low confidence |
| 150 | GPT-OSS 20B | 46.6% | estimated ± 5.3 pp, low confidence |
| 151 | Trinity-Large-Preview | 46.4% | estimated ± 5.3 pp, low confidence |
| 152 | Trinity-Large-Thinking | 46.4% | estimated ± 5.3 pp, low confidence |
| 153 | Nemotron 3 Nano 30B | 46.3% | estimated ± 5.3 pp, low confidence |
| 154 | DeepSeek V3 | 46.2% | estimated ± 5.3 pp, low confidence |
| 155 | Celeris-1 | 46.1% | estimated ± 5.3 pp, low confidence |
| 156 | Llama 4 Maverick | 46.0% | estimated ± 5.3 pp, low confidence |
| 157 | Llama 4 Scout | 46.0% | estimated ± 5.3 pp, low confidence |
| 158 | GPT-5 | 45.9% | estimated ± 6.0 pp, low confidence |
| 159 | Gemma 3 27B | 45.7% | estimated ± 5.3 pp, low confidence |
| 160 | GPT-5.1-Codex | 44.9% | estimated ± 9.3 pp, low confidence |
| 161 | Grok Build 0.1 | 44.2% | estimated ± 9.3 pp, low confidence |
| 162 | Muse Spark | 43.5% | measured |
| 163 | Claude Sonnet 4.5 | 43.2% | estimated ± 9.3 pp, low confidence |
| 164 | GLM-5 | 43.2% | measured |
| 165 | Ornith-1.5-9B | 42.2% | estimated ± 7.3 pp, low confidence |
| 166 | Qwen3.5 Plus | 42.2% | estimated ± 11.4 pp, low confidence |
| 167 | Grok 4.1 Fast | 41.6% | estimated ± 9.3 pp, low confidence |
| 168 | MiMo-V2.5 | 41.0% | estimated ± 9.3 pp, low confidence |
| 169 | Kimi K2.5 | 39.0% | estimated ± 6.0 pp, low confidence |
| 170 | Spark-X2.6-4B | 38.9% | estimated ± 7.3 pp, low confidence |
| 171 | Claude Haiku 4.5 | 37.9% | estimated ± 11.4 pp, low confidence |
| 172 | GPT-5.1 | 37.7% | estimated ± 6.0 pp, low confidence |
| 173 | Qwen3 Max | 36.7% | estimated ± 9.3 pp, low confidence |
| 174 | K-EXAONE 2.0 | 36.0% | estimated ± 6.0 pp, low confidence |
| 175 | Sarvam 105B | 35.4% | estimated ± 7.3 pp, low confidence |
| 176 | Grok 4 | 34.8% | estimated ± 9.3 pp, low confidence |
| 177 | LongCat-Flash-Lite-Sparse | 34.5% | estimated ± 7.3 pp, low confidence |
| 178 | Claude 4 Sonnet | 31.4% | estimated ± 9.3 pp, low confidence |
| 179 | Gemini 3.1 Flash-Lite | 29.9% | estimated ± 9.3 pp, low confidence |
| 180 | Grok 4.20 | 29.7% | estimated ± 9.3 pp, low confidence |
| 181 | MiMo-V2-Pro | 27.7% | estimated ± 9.3 pp, low confidence |
| 182 | MiniCPM5-2B | 25.9% | estimated ± 6.0 pp, low confidence |
| 183 | Sarvam 30B | 23.2% | estimated ± 7.3 pp, low confidence |
| 184 | GLM-5V-Turbo | 20.6% | estimated ± 9.3 pp, low confidence |
| 185 | DeepSeek V3.2 | 19.3% | estimated ± 9.3 pp, low confidence |
| 186 | MiMo-V2-Flash | 16.8% | estimated ± 6.0 pp, low confidence |
| 187 | Granite 4.2 30B | 10.0% | estimated ± 6.0 pp, low confidence |
| 188 | Gemma 4 26B A4B | 8.8% | estimated ± 6.0 pp, low confidence |
| 189 | Ling 3.0 Tiny | 8.2% | estimated ± 6.0 pp, low confidence |
| 190 | Claude 3.5 Haiku | 0.0% | estimated ± 6.0 pp, low confidence |
| 191 | DeepSeek V3 0324 | 0.0% | estimated ± 6.0 pp, low confidence |
| 192 | Gemma 4 12B | 0.0% | estimated ± 6.0 pp, low confidence |
| 193 | Gemma 4 E2B | 0.0% | estimated ± 6.0 pp, low confidence |
| 194 | Gemma 4 E4B | 0.0% | estimated ± 6.0 pp, low confidence |
| 195 | GPT-4.1 mini | 0.0% | estimated ± 6.0 pp, low confidence |
| 196 | GPT-4.1 nano | 0.0% | estimated ± 6.0 pp, low confidence |
| 197 | GPT-4o | 0.0% | estimated ± 6.0 pp, low confidence |
| 198 | GPT-4o mini | 0.0% | estimated ± 6.0 pp, low confidence |
| 199 | Granite 4.2 3B | 0.0% | estimated ± 6.0 pp, low confidence |
| 200 | K-Exaone | 0.0% | estimated ± 6.0 pp, low confidence |
| 201 | LFM2.5-2.6B | 0.0% | estimated ± 6.0 pp, low confidence |
| 202 | Ling 2.6 Flash | 0.0% | estimated ± 6.0 pp, low confidence |
| 203 | Llama 3.3 70B | 0.0% | estimated ± 6.0 pp, low confidence |
| 204 | Mercury 2.5 | 0.0% | estimated ± 6.0 pp, low confidence |
| 205 | Mistral Small 3.1 | 0.0% | estimated ± 6.0 pp, low confidence |
| 206 | Nemotron 3 Nano Omni 30B A3B | 0.0% | estimated ± 6.0 pp, low confidence |
| 207 | North Mini Code | 0.0% | estimated ± 6.0 pp, low confidence |
| 208 | Solar Pro 3 | 0.0% | estimated ± 6.0 pp, low confidence |
| 209 | Ultravox v0.6 Llama 3.3 70B | 0.0% | estimated ± 6.0 pp, low confidence |
All leaderboards
Agentic · terminal
- AA Terminal-Bench 2.1
- AA Terminal-Bench 4.0
- Terminal-Bench 4.0
- Terminal-Bench 2.0
- Terminal-Bench 2.1
- Terminal-Bench 3.0
- Terminal-Bench Hard
- Terminal-Bench-Science 0.1
- Terminal-Bench 4.1
- Terminal-Bench 2.1 (Vals AI)
Agentic · tools
- AA Agentic Index
- AA-AnalystAgent
- AA AutomationBench
- CWE-Bench-AA
- AA Cyber Index
- DeepsecBench-AA
- AA EnterpriseOps-Gym
- GDP.pdf
- AA Harvey LAB v1.0
- Harvey LAB-AA v1.1
- AA ITBench
- AA Tau3 Banking
- Agents' Last Exam
- APEX-Agents
- APEX-Agents-AA
- ApprenticeBench
- AutomationBench
- BFCL v4
- BrowseComp
- Claw-Eval
- CWE-bench v1
- CWE-bench
- CyberGym
- DeepPlanning
- DeepSearchQA
- DRACO
- ExploitGym
- Finance Agent v2
- GDPval-AA
- Gert Labs
- Hermes Bench
- Hermes Index
- HLE w/ tools
- JobBench
- LifeVal
- MazeBench
- MCP Atlas
- MCP-Tasks
- MKA
- OSWorld 2.1
- OSWorld 2.1 Offline
- OSWorld-Verified
- OSWorld 2.0
- PinchBench
- QwenClawBench
- ResearchClawBench
- SkillsBench
- τ²-bench results
- τ³-bench results
- Toolathlon
- Toolathlon-Verified
- VITA-Bench
- WideResearch
Coding
- AA Coding Index
- AA LiveCodeBench
- AA-SciCode
- Bug Hunt Bench
- CursorBench 3.1
- CursorBench 3.2
- CursorBench 4.0
- cursorBench41
- DeepSWE
- FrontierCode 1.1 Main
- FrontierCode 2.0
- FrontierCode 1.1 Extended
- FrontierSWE v2
- LiveCodeBench
- LiveCodeBench Pro
- LiveCodeBench v6
- NL2Repo
- OpenHarmony Bench v1.2
- OpenHarmony Bench v1.0
- PostTrainBench
- PostTrainBench v1
- PostTrainBench v1.1
- ProgramBench
- React Native Evals
- ScarfBench
- SciCode
- SWE-Atlas Codebase QnA
- SWE-bench Pro
- SWE-bench Verified
- SWE-Rebench
- SWE-Marathon
- Vals LiveCodeBench
- Vals SWE-bench
- Vibe Code Bench
- VulcanBench Frontier v4
- VulcanBench v3
Math
- AIME 2024
- AIME 2025
- AIME26
- FrontierMath v2 (Tier 4)
- FrontierMath v2 (Tiers 1-3)
- HMMT Feb 2025
- HMMT Feb 2026
- HMMT Nov 2025
- IMOAnswerBench
- MATH-500
- MMAnswerBench
Knowledge & reasoning
- AA-GPQA Diamond
- HealthBench Professional (length-adjusted, Artificial Analysis)
- HealthBench Professional (raw, Artificial Analysis)
- AA-HLE
- AA MMLU-Pro
- ARC-AGI-2
- ARC-AGI-3 Provider Adapter
- ARC-AGI-1
- ARC-AGI-3
- Artificial Analysis Intelligence Index
- BBH
- BioBasicsBench
- BioMysteryBench (human-difficult)
- BioMysteryBench (human-solvable)
- C-Eval
- CritPt
- GPQA
- GPQA Diamond
- HealthBench (raw)
- HealthBench Hard
- HealthBench (length-adjusted)
- HealthBench Professional
- HealthBench Professional (raw)
- HLE
- HLE w/o tools
- HLE-Verified
- JevBench 1.4
- LABBench2
- MedXpertQA (Text)
- MMLU
- MMLU-Pro
- MMLU-Pro (Arcee)
- MMLU-Redux
- MMMLU
- AA-Omniscience Accuracy
- SimpleQA
- STEMBench
- SuperGPQA
- Vals GPQA Diamond
- Vals MMLU-Pro
Instruction following
Multilingual
Vision & documents
- AA-MMMU-Pro
- AI2D_TEST
- BabyVision
- BabyVision w/ Python
- BenchCAD Vision2Code (no tools)
- BenchCAD Vision2Code (tools)
- CC-OCR
- Chartography (no tools)
- Chartography (tools)
- CharXiv
- CharXiv w/o tools
- CountBench
- DynaMath
- ERQA
- LVBench
- MathVision
- MathVision w/ Python
- MedXpertQA (MM)
- MLVU (M-Avg)
- MMMU
- MMMU-Pro
- MMMU-Pro w/ Python
- MMVU
- OCRBench V2
- OfficeQA
- OfficeQA Pro
- OmniDocBench 1.5
- ParseBench-1
- ParseBench-2
- PerceptionBench
- RealWorldQA
- RefCOCO (avg)
- ScreenSpot Pro
- SimpleVQA
- SWE-bench Multimodal
- Video-MME (w/o subtitle)
- Video-MME (with subtitle)
- VideoMMMU
- V*
- We-Math
- ZeroBench
- ZeroBench w/ Python
Long context
General
- HLE-MultiChoice (Decision Index)
- Artificial Analysis Decision Index v1.0
- Artificial Analysis Decision Index
- When2Call (Decision Index)
- WildGuardTest (Decision Index)
- BabaIsBench
- BinaryAudit
- BlenderWorld (text-only)
- Chess Puzzles
- CompileBench
- Cookie-Bench
- DAYJOB
- DAYJOB Healthcare
- Decision Index Full Score
- EBR-bench
- ExploitBench
- ExploitBench Internal Port
- FACTS Multimodal
- FailureSensorIQ
- FrontierPhysics
- Furniture Assembly
- Decision Index Arts
- Decision Index Knowledge
- Decision Index Language
- Decision Index Retrieval
- Decision Index Tools
- MaxxBench
- MirrorCode
- Mystery Game Puzzles
- NYT Connections
- OTelBench
- SEC-Bench Pro
- SkillsBench official leaderboard (BenchFlow)
- SpireClimb
- StealthMark