Long context
AA-LCR leaderboard
As of 2026-11-21, the highest measured score on AA-LCR is 88.7% by Kimi K3. 12 more models have estimated scores, calibrated from the benchmarks they were measured on.
Measured scores: benchlm.ai.
| # | Model | Score | Source |
|---|---|---|---|
| 1 | Kimi K3 | 88.7% | measured |
| 2 | Step 5 Preview | 88.3% | measured |
| 3 | Kimi K3.1 Pro | 86.3% | measured |
| 4 | MiMo-V2.6-Pro | 86.3% | measured |
| 5 | Claude Fable 5.1 | 85.3% | measured |
| 6 | Claude Opus 5.5 | 84.7% | measured |
| 7 | GPT-5.5 | 84.3% | measured |
| 8 | DeepSeek V4.1 Flash 1017 | 84.0% | measured |
| 9 | GPT-5.6 Sol | 84.0% | measured |
| 10 | GPT-5.6 Luna | 83.7% | measured |
| 11 | GPT-6 Sol | 83.7% | measured |
| 12 | Muse Wonder Full | 83.7% | measured |
| 13 | Claude Fable 5.2 | 83.3% | measured |
| 14 | GPT-5.3 Codex | 83.3% | measured |
| 15 | GPT-6 Luna | 83.3% | measured |
| 16 | Inkling 2 Small | 83.3% | measured |
| 17 | Muse Glimmer 30B | 83.3% | measured |
| 18 | GPT-5.6 Terra | 83.0% | measured |
| 19 | GPT-6.1 Sol | 83.0% | measured |
| 20 | Ling 3.1 Flash | 83.0% | measured |
| 21 | MiniMax M3 | 83.0% | measured |
| 22 | Muse Spark 1.3 | 83.0% | measured |
| 23 | Claude Haiku 5.5 | 82.7% | measured |
| 24 | Claude Sonnet 5.5 | 82.7% | measured |
| 25 | GPT-5.2 | 82.7% | measured |
| 26 | GPT-6.1 Terra | 82.7% | measured |
| 27 | Claude Fable 5 | 82.3% | measured |
| 28 | GPT-5.2-Codex | 82.3% | measured |
| 29 | Claude Sonnet 5 | 82.0% | measured |
| 30 | Gemini 3.1 Pro | 82.0% | measured |
| 31 | GPT-5.4 | 82.0% | measured |
| 32 | Qwen3.8-27B | 82.0% | measured |
| 33 | Gemini 3.7 Flash | 81.7% | measured |
| 34 | Gemini 3.8 Flash | 81.3% | measured |
| 35 | Kimi K2.6 | 81.0% | measured |
| 36 | Sage-A1.5 | 81.0% | measured |
| 37 | GPT-6 Astra | 80.7% | measured |
| 38 | Muse Wonder | 80.7% | measured |
| 39 | Qwen 3.6 Max (preview) | 80.7% | measured |
| 40 | DeepSeek V4 Pro | 80.3% | measured |
| 41 | Grok 4.6 | 80.3% | measured |
| 42 | Qwen3.8 Max Preview | 80.3% | measured |
| 43 | Gemini 3.6 Flash | 80.0% | measured |
| 44 | GPT-5.1 | 80.0% | measured |
| 45 | DeepSeek V4 Flash 0731 | 79.7% | measured |
| 46 | Gemini 4 Argon | 79.7% | measured |
| 47 | GLM-5.3 | 79.7% | measured |
| 48 | MiMo-V2.5-Pro | 79.7% | measured |
| 49 | MiniMax M3.1 Flash | 79.7% | measured |
| 50 | Qwen3.8-Flash-Next | 79.7% | measured |
| 51 | Inkling 2 Mini | 79.5% | estimated ± 12.5 pp, low confidence |
| 52 | Beam 1.0 501B-A23B FP8 | 79.4% | measured |
| 53 | Apodex 1.1 | 79.3% | measured |
| 54 | Apodex 1.1 Mini | 79.3% | measured |
| 55 | Claude Opus 5 | 79.3% | measured |
| 56 | Gemini 4 Carbon | 79.3% | measured |
| 57 | Grok 4.5 | 79.3% | measured |
| 58 | Kimi K2.7 Code | 79.3% | measured |
| 59 | Hy3 | 79.0% | measured |
| 60 | Muse Spark 1.2 | 79.0% | measured |
| 61 | Qwen3.7 Max | 79.0% | measured |
| 62 | Claude Opus 4.7 | 78.7% | measured |
| 63 | Beam 1.0 501B-A23B NVFP4 | 78.7% | measured |
| 64 | GLM-5.2 | 78.3% | measured |
| 65 | Ling 3.0 Flash VL | 78.3% | measured |
| 66 | MiniMax M2.7 | 78.3% | measured |
| 67 | Muse Spark Open | 78.3% | measured |
| 68 | Qwen3.6 Plus | 78.3% | measured |
| 69 | GPT-5 | 78.2% | measured |
| 70 | Beam 1.0 501B-A23B | 78.2% | measured |
| 71 | Claude Opus 4.6 (Adaptive) | 78.0% | measured |
| 72 | Kimi K2.5 | 78.0% | measured |
| 73 | Muse Spark | 78.0% | measured |
| 74 | Claude Opus 4.8 | 77.7% | measured |
| 75 | Muse Spark 1.1 | 77.7% | measured |
| 76 | Qwen3.5-27B | 77.7% | measured |
| 77 | Claude Opus 4.5 Thinking | 77.3% | measured |
| 78 | Inkling | 77.3% | measured |
| 79 | Qwen3.5 397B | 77.3% | measured |
| 80 | Qwen3.6-27B | 77.3% | measured |
| 81 | GPT-5.4 mini | 77.0% | measured |
| 82 | Ternary Bonsai 2 27B | 77.0% | measured |
| 83 | GPT-5.4 nano | 76.7% | measured |
| 84 | Grok 4.7 | 76.7% | measured |
| 85 | Quasar 438B | 76.3% | measured |
| 86 | Qwen3.5-122B-A10B | 76.3% | measured |
| 87 | Gemini 3.5 Flash-Lite | 76.0% | measured |
| 88 | Gemini 3 Pro | 76.0% | measured |
| 89 | GLM-5 | 75.7% | measured |
| 90 | Inkling-Small | 75.7% | measured |
| 91 | MiMo-V2-Omni | 75.0% | measured |
| 92 | Ripple-65B Preview | 75.0% | measured |
| 93 | o3 | 74.7% | measured |
| 94 | Sakana Fugu-Ultra | 74.3% | estimated ± 3.4 pp, medium confidence |
| 95 | MiMo-V2.6-Flash | 74.3% | measured |
| 96 | Qwen3.8 Max | 74.3% | estimated ± 3.4 pp, medium confidence |
| 97 | Sakana Fugu | 74.1% | estimated ± 3.4 pp, high confidence |
| 98 | Grok 4.1 Fast (Reasoning) | 74.0% | measured |
| 99 | GLM-5.1 | 73.7% | measured |
| 100 | Grok 4 Fast (Reasoning) | 73.7% | measured |
| 101 | Step 3.7 Flash | 73.7% | measured |
| 102 | Mercury 3 | 73.5% | estimated ± 3.4 pp, high confidence |
| 103 | Ling 3.0 Flash | 73.0% | measured |
| 104 | Qwen3.7 Plus | 73.0% | measured |
| 105 | Qwen3.5-35B-A3B | 72.0% | measured |
| 106 | GLM-5-Turbo | 71.7% | measured |
| 107 | Qwen3.6-35B-A3B | 71.7% | measured |
| 108 | Pokee-Isaac 28B | 71.6% | estimated ± 3.4 pp, high confidence |
| 109 | GLM-4.7 | 71.0% | measured |
| 110 | Solar Pro 4 | 71.0% | measured |
| 111 | Claude Opus 4.5 | 70.7% | measured |
| 112 | Ling 3.0 Flash Sante | 70.5% | measured |
| 113 | GLM-5V-Turbo | 70.3% | measured |
| 114 | MiniMax M1 80k | 70.3% | estimated ± 12.5 pp, low confidence |
| 115 | Gemma 4 31B | 69.7% | measured |
| 116 | Nemotron 3.5 Bolt 74B-A9B | 69.7% | measured |
| 117 | Claude 4 Opus Thinking | 69.3% | measured |
| 118 | Gemini 3.5 Flash | 69.3% | measured |
| 119 | GPT-5.1-Codex | 69.3% | measured |
| 120 | GPT-5.1-Codex-Max | 69.3% | measured |
| 121 | Mistral Medium 3.5 128B | 69.3% | measured |
| 122 | Gemini 2.5 Pro | 69.0% | measured |
| 123 | Nemotron 3.5 Super VL 120B A12B | 68.6% | measured |
| 124 | Claude Sonnet 4.6 | 68.3% | measured |
| 125 | GPT-4.1 | 68.3% | measured |
| 126 | MiMo-V2-Pro | 68.3% | measured |
| 127 | Agents-A1 | 68.1% | estimated ± 12.5 pp, low confidence |
| 128 | Grok 4 | 68.0% | measured |
| 129 | Mercury 2.5 | 68.0% | measured |
| 130 | Nemotron 3.5 Super VL 120B A12B NVFP4 | 67.7% | measured |
| 131 | Nemotron 3.5 Bolt 74B-A9B BF16 | 67.4% | measured |
| 132 | Spark-X2.6-4B | 67.3% | measured |
| 133 | Claude Opus 4.6 | 67.0% | measured |
| 134 | Nemotron 3 Ultra | 67.0% | measured |
| 135 | Hy3 Preview | 66.7% | measured |
| 136 | A.X K2 | 66.0% | measured |
| 137 | Gemma 4 26B A4B | 65.7% | measured |
| 138 | o1 | 65.0% | measured |
| 139 | Grok 4.3 | 64.3% | measured |
| 140 | Gemma 4 12B | 63.7% | measured |
| 141 | Qwen3.5-4B | 63.0% | measured |
| 142 | Solar Open 2 | 62.3% | measured |
| 143 | K-Exaone | 61.3% | measured |
| 144 | o4-mini | 61.0% | measured |
| 145 | Ling 3.0 Tiny | 60.3% | measured |
| 146 | MiniCPM5-2B | 59.0% | measured |
| 147 | K-EXAONE 2.0 | 56.2% | measured |
| 148 | Agents-A1-4B | 55.7% | estimated ± 12.5 pp, low confidence |
| 149 | DeepSeek-R1 | 55.7% | measured |
| 150 | Gemini 3 Flash | 55.3% | measured |
| 151 | Nemotron 3.5 Bolt E4B BF16 | 53.5% | estimated ± 12.5 pp, low confidence |
| 152 | Grok Code Fast 1 | 53.0% | measured |
| 153 | Kimi K2 | 53.0% | measured |
| 154 | LFM3-1.7B | 52.8% | measured |
| 155 | Command A+ | 52.7% | measured |
| 156 | GPT-OSS 120B | 52.0% | measured |
| 157 | Claude 3.7 Sonnet | 51.7% | measured |
| 158 | Nemotron 3.5 Bolt E4B | 50.9% | estimated ± 12.5 pp, low confidence |
| 159 | Llama 4 Maverick | 50.0% | measured |
| 160 | Qwen3 Max | 50.0% | measured |
| 161 | Gemini 2.5 Flash | 49.9% | measured |
| 162 | Mistral Small 4 | 49.7% | measured |
| 163 | GPT-4o | 49.3% | measured |
| 164 | Nemotron 3.5 Lightning 30B A3B NVFP4 | 49.2% | measured |
| 165 | Granite 4.2 30B | 49.0% | measured |
| 166 | DeepSeek V3.1 | 47.0% | measured |
| 167 | GLM-4.5-Air | 46.7% | measured |
| 168 | DeepSeek V3.2 | 45.7% | measured |
| 169 | Granite 4.2 8B | 45.0% | measured |
| 170 | Claude 4 Sonnet | 44.0% | measured |
| 171 | GPT-4.1 mini | 44.0% | measured |
| 172 | DeepSeek V3 0324 | 40.7% | measured |
| 173 | Nemotron 3 Nano Omni 30B A3B | 39.7% | measured |
| 174 | Celeris-1 | 38.0% | measured |
| 175 | Trinity-Large-Preview | 38.0% | measured |
| 176 | Trinity-Large-Thinking | 38.0% | measured |
| 177 | North Mini Code | 37.3% | measured |
| 178 | MiMo-V2-Flash | 36.3% | measured |
| 179 | Mistral Large 3 | 36.0% | measured |
| 180 | GPT-OSS 20B | 34.7% | measured |
| 181 | LLaDA2.2-mini | 33.6% | estimated ± 12.5 pp, low confidence |
| 182 | Solar Pro 3 | 32.3% | measured |
| 183 | Gemini 2.5 Flash-Lite | 32.0% | measured |
| 184 | Gemma 4 E4B | 32.0% | measured |
| 185 | Gemini 2.0 Flash | 31.3% | measured |
| 186 | Grok 4.1 Fast | 31.3% | measured |
| 187 | Ling 2.6 Flash | 31.3% | measured |
| 188 | Mistral Medium 3 | 31.3% | measured |
| 189 | DeepSeek V3 | 29.3% | measured |
| 190 | Claude 3 Haiku | 27.7% | measured |
| 191 | Llama 4 Scout | 27.7% | measured |
| 192 | Claude 3.5 Haiku | 27.3% | measured |
| 193 | QwQ-32B | 26.7% | measured |
| 194 | GLM-4.6 | 26.3% | measured |
| 195 | Llama 3.1 405B | 25.3% | measured |
| 196 | Granite 4.2 3B | 24.3% | measured |
| 197 | Mistral Small 3.1 | 22.3% | measured |
| 198 | Nova Pro | 21.0% | measured |
| 199 | GPT-4.1 nano | 20.3% | measured |
| 200 | Devstral Small | 18.7% | measured |
| 201 | Gemma 4 E2B | 16.3% | measured |
| 202 | Llama 3.3 70B | 15.7% | measured |
| 203 | Ultravox v0.6 Llama 3.3 70B | 15.7% | measured |
| 204 | Nemotron 3 Nano 30B | 10.7% | measured |
| 205 | DeepSeek R1 Distill Qwen 32B | 8.7% | measured |
| 206 | Gemma 3 27B | 7.3% | measured |
| 207 | Granite-4.0-H-1B | 7.0% | measured |
| 208 | LFM2.5-2.6B | 5.7% | measured |
| 209 | Llama 3.1 70B | 4.7% | measured |
| 210 | Exaone 4.0 1.2B | 0.0% | measured |
| 211 | Granite-4.0-350M | 0.0% | measured |
| 212 | Granite-4.0-H-350M | 0.0% | measured |
| 213 | LFM2.5-8B-A1B | 0.0% | measured |
| 214 | LFM2.5-VL-1.6B-Extract | 0.0% | measured |
| 215 | Magistral Medium | 0.0% | measured |
| 216 | Magistral Small | 0.0% | measured |
| 217 | Phi-4 | 0.0% | measured |
| 218 | Phi-4-reasoning | 0.0% | measured |
| 219 | Qwen3-Omni-30B-A3B-Instruct | 0.0% | measured |
| 220 | Qwen3-Omni-30B-A3B-Thinking | 0.0% | measured |
| 221 | Sarvam 105B | 0.0% | measured |
| 222 | Sarvam 30B | 0.0% | measured |
| 223 | Solar Pro 2 | 0.0% | measured |
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