score · Higher is better 128 results
32.3% OpenAI GPT-5.6 Sol max effort 31.1% Claude Claude Fable 5.1 xhigh effort 30.3% Claude Claude Fable 5.1 high effort 30.0% OpenAI GPT-5.6 Terra max effort 29.7% Claude Claude Fable 5.1 max effort 29.1% Claude Claude Fable 5.1 medium effort 29.1% Claude Claude Opus 5 max effort 28.6% OpenAI GPT-5.6 Sol xhigh effort 28.6% Claude Claude Fable 5 max effort 28.3% Claude Claude Opus 5 high effort 27.7% Claude Claude Fable 5.1 low effort 27.7% Claude Claude Opus 5 xhigh effort 27.1% OpenAI GPT-5.6 Terra xhigh effort 27.1% OpenAI GPT-5.5 xhigh effort 26.9% Claude Claude Opus 5 medium effort 26.0% Meta Muse Spark 1.3 xhigh effort 25.7% OpenAI GPT-5.6 Sol high effort 25.4% OpenAI GPT-5.5 high effort 24.9% Meta Muse Spark 1.3 max effort 23.4% OpenAI GPT-5.4 xhigh effort 23.4% Kimi Kimi K3 max effort 23.1% Claude Claude Opus 5 low effort 22.9% OpenAI GPT-5.6 Sol medium effort 22.9% OpenAI GPT-5.6 Terra high effort 20.9% Claude Claude Opus 4.8 max effort 20.9% Zhipu GLM-5.2 max effort 20.6% OpenAI GPT-5.6 Luna max effort 20.6% OpenAI GPT-5.6 Luna xhigh effort 20.0% Qwen Qwen3.8 Max xhigh effort 20.0% Qwen Qwen3.8 2.4T A95B xhigh effort 19.7% Grok Grok 4.6 xhigh effort 19.1% Zhipu GLM-5.3 max effort 18.6% OpenAI GPT-5.5 medium effort 18.3% Gemini Gemini 3.8 Flash high effort 18.0% DeepSeek DeepSeek V4 Pro max effort 17.7% Meta Muse Spark 1.2 xhigh effort 17.7% Grok Grok 4.6 medium effort 17.4% OpenAI GPT-5.6 Terra medium effort 17.1% Grok Grok 4.6 high effort 16.9% Claude Claude Sonnet 5 max effort 16.6% DeepSeek DeepSeek V4 Flash max effort 16.6% OpenAI GPT-5.6 Luna high effort 15.4% Grok Grok 4.5 high effort 15.4% Zhipu GLM-5.3 Flash max effort 14.9% OpenAI GPT-5.6 Sol low effort 14.3% Gemini Gemini 3.7 Flash high effort 13.4% Qwen Qwen3.7 Max unspecified effort 12.3% Gemini Gemini 3.8 Flash medium effort 10.6% Gemini Gemini 3.6 Flash high effort 10.0% Kimi Kimi K2.7 Code enabled effort 9.4% Gemini Gemini 3.7 Flash medium effort 9.4% OpenAI GPT-5.6 Terra low effort 8.0% Kimi Kimi K2.6 enabled effort 8.0% OpenAI GPT-5.5 low effort 8.0% Grok Grok 4.3 high effort 7.4% OpenAI GPT-5.4 low effort 5.7% Gemini Gemini 3.7 Flash low effort 5.7% Grok Grok 4.6 low effort 5.4% Qwen Qwen3.8 27B xhigh effort 5.4% TM Inkling max effort 5.1% OpenAI GPT-5.6 Sol none effort 4.9% OpenAI GPT-5.6 Luna medium effort 4.9% Tencent Tencent Hy3 high effort 4.9% Grok Grok 4.3 medium effort 4.6% Zhipu GLM-5.1 enabled effort 4.3% XiaomiMiMo MiMo V2 Flash enabled effort 4.0% Gemini Gemini 3.8 Flash low effort 4.0% XiaomiMiMo MiMo V2.5 Pro enabled effort 3.7% Minimax MiniMax M3 adaptive effort 3.7% XiaomiMiMo MiMo V2.5 enabled effort 3.4% DeepSeek DeepSeek V4 Flash high effort 3.1% Kimi Kimi K3 low effort 3.1% Kimi Kimi K2.5 enabled effort 3.1% Nvidia NVIDIA Nemotron 3 Ultra enabled effort 3.1% Nvidia NVIDIA Nemotron 3 Super enabled effort 3.1% Zhipu GLM-5.2 disabled effort 2.9% DeepSeek DeepSeek V3.2 enabled effort 2.6% Meta Muse Glimmer 30B high effort 2.6% OpenAI GPT-5.6 Luna low effort 2.5% StepFun Step 3.5 Flash Reasoning enabled 2.3% StepFun Step 3.7 Flash high effort 2.0% OpenAI GPT-5.6 Terra none effort 1.4% Gemini Gemma 4 31B IT enabled effort 1.4% Kimi Kimi K2.6 disabled effort 1.4% OpenAI GPT-5.5 none effort 1.4% OpenAI GPT-OSS 20B high effort 1.1% Minimax MiniMax M2.5 Reasoning enabled 1.1% OpenAI GPT-OSS 120B high effort 1.1% Qwen Qwen3.6 27B enabled effort 1.1% XiaomiMiMo MiMo V2.5 Pro disabled effort 0.9% DeepSeek DeepSeek V3.2 disabled effort 0.9% Qwen Qwen3.6 27B disabled effort 0.6% DeepSeek DeepSeek R1 Published default 0.6% Minimax MiniMax M2.7 Reasoning enabled 0.6% Kimi Kimi K2.5 disabled effort 0.6% OpenAI GPT-5.4 none effort 0.6% Grok Grok 4.3 low effort 0.3% Nova Amazon Nova 2 Lite High run 0.3% Cohere North Mini Code 1.0 Reasoning enabled 0.3% Cohere Command A+ 05-2026 Reasoning enabled 0.3% Mistral Mistral Small 4 high effort 0.3% Mistral Mistral Small 4 none effort 0.3% OpenAI GPT-5.6 Luna none effort 0.3% Qwen Qwen3.6 35B A3B enabled effort 0.0% AI21 Jamba Reasoning 3B Reasoning enabled 0.0% AI21 Jamba Large 1.7 Non-reasoning run 0.0% Nova Amazon Nova Pro Published default 0.0% Nova Amazon Nova 2 Lite Low run 0.0% Nova Amazon Nova 2 Lite Medium run 0.0% Nova Amazon Nova 2 Lite No reasoning 0.0% Nova Amazon Nova Premier Non-reasoning run 0.0% Baidu ERNIE 4.5 300B A47B Non-reasoning run 0.0% Cohere Tiny Aya Global Non-reasoning run 0.0% Gemini Gemma 4 31B IT disabled effort 0.0% Meta Llama 4 Maverick 17B 128E Instruct Non-reasoning run 0.0% Meta Llama 4 Scout 17B 16E Instruct Non-reasoning run 0.0% Mistral Mistral Large 3 Published default 0.0% Mistral Mistral Medium 3.5 high effort 0.0% Nvidia Nemotron 3 Nano Omni 30B A3B Reasoning enabled effort 0.0% OpenAI GPT-OSS 120B low effort 0.0% OpenAI GPT-OSS 20B low effort 0.0% Qwen Qwen3.6 35B A3B disabled effort 0.0% Qwen Qwen3.8 27B low effort 0.0% Qwen Qwen3.8 27B medium effort 0.0% StepFun Step3-VL 10B Reasoning enabled 0.0% Grok Grok 4.3 none effort 0.0% XiaomiMiMo MiMo V2 Flash disabled effort 0.0% Zhipu GLM-5.1 disabled effort index · Higher is better 128 results
71.7% Claude Claude Fable 5.1 max effort 71.7% Claude Claude Fable 5 max effort 71.2% Claude Claude Fable 5.1 xhigh effort 70.4% Claude Claude Fable 5.1 high effort 68.8% Claude Claude Fable 5.1 medium effort 68.5% Claude Claude Opus 5 max effort 67.7% Claude Claude Opus 5 xhigh effort 67.1% Claude Claude Fable 5.1 low effort 66.9% Claude Claude Opus 5 high effort 65.5% Claude Claude Opus 5 medium effort 65.2% Grok Grok 4.6 high effort 64.8% Gemini Gemini 3.8 Flash high effort 64.7% Grok Grok 4.6 xhigh effort 64.4% Claude Claude Opus 4.8 max effort 64.3% Gemini Gemini 3.8 Flash medium effort 64.3% Claude Claude Opus 5 low effort 64.0% Grok Grok 4.6 medium effort 63.6% Meta Muse Spark 1.2 xhigh effort 63.2% Gemini Gemini 3.7 Flash high effort 63.0% Grok Grok 4.6 low effort 62.7% Grok Grok 4.5 high effort 62.5% Meta Muse Spark 1.3 max effort 61.9% Gemini Gemini 3.7 Flash medium effort 61.6% Meta Muse Spark 1.3 xhigh effort 61.1% Gemini Gemini 3.6 Flash high effort 61.1% Gemini Gemini 3.7 Flash low effort 61.0% OpenAI GPT-5.6 Sol max effort 60.7% Gemini Gemini 3.8 Flash low effort 60.5% OpenAI GPT-5.6 Sol xhigh effort 60.3% OpenAI GPT-5.5 xhigh effort 60.2% OpenAI GPT-5.6 Sol high effort 59.9% Kimi Kimi K3 max effort 59.7% OpenAI GPT-5.6 Sol medium effort 59.4% OpenAI GPT-5.6 Sol low effort 59.4% OpenAI GPT-5.5 high effort 59.1% OpenAI GPT-5.5 medium effort 59.0% Grok Grok 4.3 high effort 58.4% Grok Grok 4.3 medium effort 58.2% Claude Claude Sonnet 5 max effort 57.6% OpenAI GPT-5.5 low effort 57.1% Zhipu GLM-5.3 max effort 57.0% Grok Grok 4.3 low effort 56.7% Qwen Qwen3.7 Max unspecified effort 53.7% Zhipu GLM-5.3 Flash max effort 52.9% OpenAI GPT-5.4 xhigh effort 52.6% Kimi Kimi K2.6 enabled effort 52.4% OpenAI GPT-5.4 low effort 52.2% Zhipu GLM-5.2 max effort 52.2% Qwen Qwen3.8 2.4T A95B xhigh effort 52.0% Kimi Kimi K3 low effort 51.7% Qwen Qwen3.8 Max xhigh effort 51.6% XiaomiMiMo MiMo V2.5 Pro enabled effort 51.0% TM Inkling max effort 50.7% Minimax MiniMax M3 adaptive effort 50.5% OpenAI GPT-5.6 Sol none effort 50.4% Zhipu GLM-5.1 enabled effort 50.4% DeepSeek DeepSeek V4 Pro max effort 50.4% Minimax MiniMax M2.7 Reasoning enabled 50.0% OpenAI GPT-5.6 Terra max effort 49.8% Nvidia NVIDIA Nemotron 3 Ultra enabled effort 48.5% OpenAI GPT-5.6 Terra xhigh effort 48.3% OpenAI GPT-5.6 Terra high effort 48.0% Cohere Command A+ 05-2026 Reasoning enabled 47.6% OpenAI GPT-5.5 none effort 47.4% OpenAI GPT-5.6 Terra medium effort 46.7% Zhipu GLM-5.2 disabled effort 46.6% OpenAI GPT-5.6 Terra low effort 46.3% Kimi Kimi K2.5 enabled effort 45.4% Kimi Kimi K2.6 disabled effort 45.1% XiaomiMiMo MiMo V2.5 enabled effort 45.0% Qwen Qwen3.8 27B xhigh effort 44.9% Kimi Kimi K2.7 Code enabled effort 44.9% OpenAI GPT-5.6 Luna max effort 44.6% OpenAI GPT-5.6 Luna xhigh effort 44.0% OpenAI GPT-5.6 Luna high effort 43.4% OpenAI GPT-5.6 Luna medium effort 43.1% Kimi Kimi K2.5 disabled effort 42.9% DeepSeek DeepSeek V4 Flash max effort 42.7% OpenAI GPT-5.6 Luna low effort 42.2% OpenAI GPT-5.4 none effort 40.8% Tencent Tencent Hy3 high effort 40.0% Qwen Qwen3.6 27B enabled effort 38.9% Qwen Qwen3.6 35B A3B enabled effort 38.8% Zhipu GLM-5.1 disabled effort 38.8% DeepSeek DeepSeek V3.2 enabled effort 38.5% DeepSeek DeepSeek V4 Flash high effort 38.4% OpenAI GPT-5.6 Terra none effort 37.5% OpenAI GPT-5.6 Luna none effort 36.7% Qwen Qwen3.8 27B low effort 34.8% Mistral Mistral Small 4 high effort 33.9% DeepSeek DeepSeek R1 Published default 33.6% Meta Muse Glimmer 30B high effort 33.5% Grok Grok 4.3 none effort 32.4% Baidu ERNIE 4.5 300B A47B Non-reasoning run 32.3% Nova Amazon Nova Premier Non-reasoning run 31.9% Qwen Qwen3.8 27B medium effort 31.6% Mistral Mistral Medium 3.5 high effort 31.4% StepFun Step 3.7 Flash high effort 31.1% XiaomiMiMo MiMo V2.5 Pro disabled effort 30.6% Minimax MiniMax M2.5 Reasoning enabled 30.2% Mistral Mistral Large 3 Published default 29.3% Nvidia NVIDIA Nemotron 3 Super enabled effort 29.1% StepFun Step 3.5 Flash Reasoning enabled 29.1% Meta Llama 4 Maverick 17B 128E Instruct Non-reasoning run 27.3% XiaomiMiMo MiMo V2 Flash enabled effort 26.6% DeepSeek DeepSeek V3.2 disabled effort 26.1% Nova Amazon Nova Pro Published default 26.0% Gemini Gemma 4 31B IT enabled effort 25.8% XiaomiMiMo MiMo V2 Flash disabled effort 25.7% Mistral Mistral Small 4 none effort 25.7% Cohere North Mini Code 1.0 Reasoning enabled 25.4% OpenAI GPT-OSS 120B high effort 24.2% Gemini Gemma 4 31B IT disabled effort 23.9% Meta Llama 4 Scout 17B 16E Instruct Non-reasoning run 23.4% Qwen Qwen3.6 27B disabled effort 23.3% OpenAI GPT-OSS 120B low effort 22.9% Nova Amazon Nova 2 Lite Low run 22.5% AI21 Jamba Large 1.7 Non-reasoning run 22.5% Nova Amazon Nova 2 Lite High run 21.5% Nova Amazon Nova 2 Lite Medium run 21.3% Nvidia Nemotron 3 Nano Omni 30B A3B Reasoning enabled effort 21.1% AI21 Jamba Reasoning 3B Reasoning enabled 20.7% OpenAI GPT-OSS 20B low effort 20.5% StepFun Step3-VL 10B Reasoning enabled 20.1% Nova Amazon Nova 2 Lite No reasoning 20.0% Qwen Qwen3.6 35B A3B disabled effort 18.5% OpenAI GPT-OSS 20B high effort 7.9% Cohere Tiny Aya Global Non-reasoning run accuracy · Higher is better 128 results
59.1% Claude Claude Fable 5.1 max effort 58.7% Claude Claude Fable 5.1 xhigh effort 55.9% Claude Claude Fable 5.1 high effort 55.5% Claude Claude Fable 5 max effort 54.9% Claude Claude Opus 5 max effort 54.4% Claude Claude Opus 5 xhigh effort 53.8% Claude Claude Fable 5.1 medium effort 52.8% Claude Claude Opus 5 high effort 51.3% Claude Claude Opus 5 medium effort 49.5% OpenAI GPT-5.6 Sol max effort 49.1% Meta Muse Spark 1.3 max effort 48.9% Claude Claude Fable 5.1 low effort 48.7% Claude Claude Opus 4.8 max effort 47.9% Gemini Gemini 3.7 Flash high effort 47.8% Gemini Gemini 3.8 Flash high effort 47.5% Meta Muse Spark 1.3 xhigh effort 47.3% OpenAI GPT-5.6 Sol xhigh effort 46.9% Kimi Kimi K3 max effort 46.0% OpenAI GPT-5.6 Sol high effort 45.8% OpenAI GPT-5.5 xhigh effort 45.5% Meta Muse Spark 1.2 xhigh effort 45.0% OpenAI GPT-5.5 high effort 44.1% Grok Grok 4.6 xhigh effort 43.7% OpenAI GPT-5.4 xhigh effort 43.4% Claude Claude Opus 5 low effort 43.0% Qwen Qwen3.8 Max xhigh effort 42.9% OpenAI GPT-5.6 Terra max effort 42.9% Grok Grok 4.6 high effort 42.7% Grok Grok 4.5 high effort 42.4% Qwen Qwen3.8 2.4T A95B xhigh effort 42.4% OpenAI GPT-5.5 medium effort 42.3% Zhipu GLM-5.3 max effort 42.2% OpenAI GPT-5.6 Sol medium effort 42.1% Gemini Gemini 3.8 Flash medium effort 42.1% Grok Grok 4.6 medium effort 41.9% OpenAI GPT-5.6 Terra xhigh effort 41.3% Claude Claude Sonnet 5 max effort 41.1% Zhipu GLM-5.2 max effort 41.0% DeepSeek DeepSeek V4 Pro max effort 40.8% Gemini Gemini 3.6 Flash high effort 40.5% Qwen Qwen3.7 Max unspecified effort 39.9% Zhipu GLM-5.3 Flash max effort 39.5% OpenAI GPT-5.6 Luna max effort 39.4% OpenAI GPT-5.6 Sol low effort 39.0% Gemini Gemini 3.7 Flash medium effort 39.0% Minimax MiniMax M3 adaptive effort 38.6% DeepSeek DeepSeek V4 Flash max effort 38.5% OpenAI GPT-5.6 Terra high effort 37.5% Kimi Kimi K2.6 enabled effort 37.2% Grok Grok 4.3 high effort 37.1% Gemini Gemini 3.8 Flash low effort 37.0% OpenAI GPT-5.6 Luna xhigh effort 35.7% XiaomiMiMo MiMo V2.5 Pro enabled effort 35.1% Gemini Gemini 3.7 Flash low effort 35.0% Kimi Kimi K2.7 Code enabled effort 33.9% Qwen Qwen3.8 27B xhigh effort 33.5% Tencent Tencent Hy3 high effort 33.4% OpenAI GPT-5.6 Luna high effort 33.3% OpenAI GPT-5.6 Terra medium effort 32.7% OpenAI GPT-5.5 low effort 31.9% TM Inkling max effort 30.8% OpenAI GPT-5.4 low effort 30.7% Kimi Kimi K2.5 enabled effort 30.3% DeepSeek DeepSeek V4 Flash high effort 30.1% Zhipu GLM-5.1 enabled effort 30.0% Grok Grok 4.3 medium effort 29.6% Minimax MiniMax M2.7 Reasoning enabled 29.2% OpenAI GPT-5.6 Terra low effort 28.4% Nvidia NVIDIA Nemotron 3 Ultra enabled effort 27.9% Zhipu GLM-5.1 disabled effort 27.6% Grok Grok 4.6 low effort 27.2% XiaomiMiMo MiMo V2.5 enabled effort 25.8% OpenAI GPT-5.6 Luna medium effort 25.0% Kimi Kimi K3 low effort 24.6% DeepSeek DeepSeek V3.2 enabled effort 23.6% Gemini Gemma 4 31B IT enabled effort 23.1% Qwen Qwen3.6 27B enabled effort 22.8% XiaomiMiMo MiMo V2 Flash enabled effort 22.2% Qwen Qwen3.6 35B A3B enabled effort 22.0% Meta Muse Glimmer 30B high effort 21.4% StepFun Step 3.7 Flash high effort 21.1% StepFun Step 3.5 Flash Reasoning enabled 20.8% Nvidia NVIDIA Nemotron 3 Super enabled effort 20.5% Minimax MiniMax M2.5 Reasoning enabled 19.8% OpenAI GPT-5.6 Luna low effort 19.6% OpenAI GPT-OSS 120B high effort 19.6% Kimi Kimi K2.6 disabled effort 18.4% Grok Grok 4.3 low effort 16.7% OpenAI GPT-5.6 Sol none effort 15.1% Qwen Qwen3.6 27B disabled effort 14.8% XiaomiMiMo MiMo V2.5 Pro disabled effort 14.1% Qwen Qwen3.8 27B medium effort 14.0% Qwen Qwen3.8 27B low effort 13.9% Qwen Qwen3.6 35B A3B disabled effort 13.8% Mistral Mistral Medium 3.5 high effort 13.7% OpenAI GPT-5.5 none effort 13.2% Kimi Kimi K2.5 disabled effort 12.0% Cohere Command A+ 05-2026 Reasoning enabled 11.8% Gemini Gemma 4 31B IT disabled effort 11.6% Nova Amazon Nova 2 Lite High run 11.4% OpenAI GPT-5.6 Terra none effort 11.3% OpenAI GPT-5.4 none effort 11.2% DeepSeek DeepSeek V3.2 disabled effort 11.1% Cohere North Mini Code 1.0 Reasoning enabled 11.0% OpenAI GPT-OSS 20B high effort 10.8% StepFun Step3-VL 10B Reasoning enabled 9.9% Mistral Mistral Small 4 high effort 9.8% Zhipu GLM-5.2 disabled effort 9.0% Nova Amazon Nova 2 Lite Medium run 8.6% XiaomiMiMo MiMo V2 Flash disabled effort 8.5% DeepSeek DeepSeek R1 Published default 7.2% OpenAI GPT-5.6 Luna none effort 6.8% Grok Grok 4.3 none effort 5.9% OpenAI GPT-OSS 120B low effort 5.3% OpenAI GPT-OSS 20B low effort 5.2% Cohere Tiny Aya Global Non-reasoning run 4.9% Meta Llama 4 Maverick 17B 128E Instruct Non-reasoning run 4.8% Nvidia Nemotron 3 Nano Omni 30B A3B Reasoning enabled effort 4.2% Nova Amazon Nova Premier Non-reasoning run 4.2% Mistral Mistral Large 3 Published default 4.0% Nova Amazon Nova 2 Lite Low run 3.8% Mistral Mistral Small 4 none effort 3.8% Meta Llama 4 Scout 17B 16E Instruct Non-reasoning run 3.8% AI21 Jamba Reasoning 3B Reasoning enabled 3.7% AI21 Jamba Large 1.7 Non-reasoning run 3.3% Baidu ERNIE 4.5 300B A47B Non-reasoning run 3.2% Nova Amazon Nova Pro Published default 2.9% Nova Amazon Nova 2 Lite No reasoning accuracy · Higher is better 128 results
95.3% Gemini Gemini 3.8 Flash high effort 94.9% Grok Grok 4.6 high effort 94.5% Gemini Gemini 3.7 Flash high effort 94.1% Meta Muse Spark 1.3 xhigh effort 94.1% OpenAI GPT-5.6 Sol max effort 93.8% Meta Muse Spark 1.3 max effort 93.7% Claude Claude Fable 5.1 max effort 93.7% Claude Claude Opus 5 high effort 93.7% Claude Claude Opus 5 xhigh effort 93.5% Gemini Gemini 3.8 Flash medium effort 93.5% Kimi Kimi K3 max effort 93.5% OpenAI GPT-5.5 xhigh effort 93.5% Qwen Qwen3.8 2.4T A95B xhigh effort 93.5% Grok Grok 4.6 medium effort 93.5% Grok Grok 4.6 xhigh effort 93.4% Claude Claude Fable 5.1 xhigh effort 93.2% Claude Claude Opus 5 max effort 93.2% OpenAI GPT-5.5 high effort 93.1% OpenAI GPT-5.6 Sol xhigh effort 93.1% Grok Grok 4.5 high effort 92.9% Minimax MiniMax M3 adaptive effort 92.8% DeepSeek DeepSeek V4 Pro max effort 92.8% Gemini Gemini 3.6 Flash high effort 92.8% OpenAI GPT-5.6 Sol high effort 92.7% Qwen Qwen3.8 Max xhigh effort 92.6% Claude Claude Fable 5 max effort 92.6% OpenAI GPT-5.6 Sol medium effort 92.6% OpenAI GPT-5.5 medium effort 92.5% OpenAI GPT-5.6 Terra max effort 92.3% Qwen Qwen3.7 Max unspecified effort 92.1% Gemini Gemini 3.7 Flash medium effort 92.0% Claude Claude Opus 4.8 max effort 92.0% Gemini Gemini 3.8 Flash low effort 92.0% OpenAI GPT-5.4 xhigh effort 91.9% Claude Claude Opus 5 medium effort 91.7% Zhipu GLM-5.3 max effort 91.2% Zhipu GLM-5.3 Flash max effort 91.1% Claude Claude Sonnet 5 max effort 91.1% Kimi Kimi K2.6 enabled effort 91.1% OpenAI GPT-5.6 Luna max effort 91.0% OpenAI GPT-5.5 low effort 90.8% DeepSeek DeepSeek V4 Flash max effort 90.8% OpenAI GPT-5.6 Terra xhigh effort 90.6% Claude Claude Fable 5.1 high effort 90.5% Qwen Qwen3.8 27B xhigh effort 90.4% Meta Muse Spark 1.2 xhigh effort 90.1% Gemini Gemini 3.7 Flash low effort 90.1% Grok Grok 4.3 high effort 89.8% OpenAI GPT-5.6 Sol low effort 89.7% Tencent Tencent Hy3 high effort 89.6% Kimi Kimi K2.7 Code enabled effort 89.6% OpenAI GPT-5.6 Terra high effort 89.5% OpenAI GPT-5.6 Luna xhigh effort 89.5% Zhipu GLM-5.2 max effort 89.2% OpenAI GPT-5.6 Luna high effort 89.0% Grok Grok 4.3 medium effort 88.9% Claude Claude Opus 5 low effort 88.6% Claude Claude Fable 5.1 medium effort 88.1% Claude Claude Fable 5.1 low effort 87.9% Kimi Kimi K2.5 enabled effort 87.9% Grok Grok 4.6 low effort 87.4% Minimax MiniMax M2.7 Reasoning enabled 87.2% OpenAI GPT-5.6 Terra medium effort 87.2% TM Inkling max effort 87.1% OpenAI GPT-5.4 low effort 86.8% Zhipu GLM-5.1 enabled effort 86.7% DeepSeek DeepSeek V4 Flash high effort 86.7% Nvidia NVIDIA Nemotron 3 Ultra enabled effort 86.6% XiaomiMiMo MiMo V2.5 Pro enabled effort 85.9% OpenAI GPT-5.6 Luna medium effort 85.7% Gemini Gemma 4 31B IT enabled effort 84.9% XiaomiMiMo MiMo V2.5 enabled effort 84.8% Minimax MiniMax M2.5 Reasoning enabled 84.6% XiaomiMiMo MiMo V2 Flash enabled effort 84.5% Qwen Qwen3.8 27B low effort 84.5% Qwen Qwen3.8 27B medium effort 84.3% OpenAI GPT-5.6 Terra low effort 84.3% Grok Grok 4.3 low effort 84.2% Kimi Kimi K3 low effort 84.2% Qwen Qwen3.6 27B enabled effort 84.1% Qwen Qwen3.6 35B A3B enabled effort 84.0% DeepSeek DeepSeek V3.2 enabled effort 83.9% Zhipu GLM-5.1 disabled effort 83.5% Meta Muse Glimmer 30B high effort 83.5% OpenAI GPT-5.6 Luna low effort 83.1% StepFun Step 3.5 Flash Reasoning enabled 82.9% Qwen Qwen3.6 27B disabled effort 81.7% Qwen Qwen3.6 35B A3B disabled effort 81.1% Nova Amazon Nova 2 Lite High run 81.1% Baidu ERNIE 4.5 300B A47B Non-reasoning run 80.9% StepFun Step 3.7 Flash high effort 80.0% Nvidia NVIDIA Nemotron 3 Super enabled effort 79.0% OpenAI GPT-5.6 Sol none effort 78.9% Kimi Kimi K2.5 disabled effort 78.8% Kimi Kimi K2.6 disabled effort 78.2% OpenAI GPT-OSS 120B high effort 76.9% Mistral Mistral Small 4 high effort 76.8% Nova Amazon Nova 2 Lite Medium run 76.8% OpenAI GPT-5.5 none effort 76.3% Gemini Gemma 4 31B IT disabled effort 76.2% XiaomiMiMo MiMo V2.5 Pro disabled effort 76.1% Cohere Command A+ 05-2026 Reasoning enabled 75.7% Cohere North Mini Code 1.0 Reasoning enabled 75.1% DeepSeek DeepSeek V3.2 disabled effort 74.8% Mistral Mistral Medium 3.5 high effort 74.8% OpenAI GPT-5.4 none effort 74.6% OpenAI GPT-5.6 Terra none effort 70.8% DeepSeek DeepSeek R1 Published default 69.8% Nova Amazon Nova 2 Lite Low run 69.0% StepFun Step3-VL 10B Reasoning enabled 68.8% OpenAI GPT-OSS 20B high effort 68.6% Zhipu GLM-5.2 disabled effort 68.0% Mistral Mistral Large 3 Published default 67.2% OpenAI GPT-OSS 120B low effort 67.1% Meta Llama 4 Maverick 17B 128E Instruct Non-reasoning run 65.8% Grok Grok 4.3 none effort 65.6% XiaomiMiMo MiMo V2 Flash disabled effort 64.5% OpenAI GPT-5.6 Luna none effort 61.1% OpenAI GPT-OSS 20B low effort 60.3% Nova Amazon Nova 2 Lite No reasoning 58.7% Meta Llama 4 Scout 17B 16E Instruct Non-reasoning run 57.1% Mistral Mistral Small 4 none effort 56.9% Nova Amazon Nova Premier Non-reasoning run 49.9% Nova Amazon Nova Pro Published default 46.9% Nvidia Nemotron 3 Nano Omni 30B A3B Reasoning enabled effort 39.0% AI21 Jamba Large 1.7 Non-reasoning run 33.3% AI21 Jamba Reasoning 3B Reasoning enabled 30.5% Cohere Tiny Aya Global Non-reasoning run score · Higher is better 127 results
88.7% Kimi Kimi K3 max effort 85.3% Claude Claude Fable 5.1 max effort 84.7% Claude Claude Fable 5.1 medium effort 84.3% Meta Muse Spark 1.3 max effort 84.3% OpenAI GPT-5.5 high effort 84.3% OpenAI GPT-5.5 xhigh effort 84.0% Gemini Gemini 3.8 Flash medium effort 84.0% OpenAI GPT-5.6 Sol max effort 83.7% Claude Claude Fable 5.1 high effort 83.7% OpenAI GPT-5.6 Luna max effort 83.3% Meta Muse Glimmer 30B high effort 83.0% Claude Claude Fable 5.1 xhigh effort 83.0% Gemini Gemini 3.7 Flash medium effort 83.0% Meta Muse Spark 1.3 xhigh effort 83.0% Minimax MiniMax M3 adaptive effort 83.0% OpenAI GPT-5.6 Terra max effort 83.0% OpenAI GPT-5.5 medium effort 82.3% Claude Claude Fable 5.1 low effort 82.3% Claude Claude Fable 5 max effort 82.3% OpenAI GPT-5.6 Sol xhigh effort 82.0% Claude Claude Opus 5 medium effort 82.0% Claude Claude Sonnet 5 max effort 82.0% OpenAI GPT-5.4 xhigh effort 82.0% Qwen Qwen3.8 27B xhigh effort 81.7% Gemini Gemini 3.7 Flash high effort 81.7% OpenAI GPT-5.6 Sol high effort 81.7% OpenAI GPT-5.6 Luna xhigh effort 81.3% Claude Claude Opus 5 low effort 81.3% Gemini Gemini 3.8 Flash high effort 81.0% Kimi Kimi K2.6 enabled effort 81.0% OpenAI GPT-5.5 low effort 81.0% Grok Grok 4.6 medium effort 81.0% Grok Grok 4.6 xhigh effort 80.7% Gemini Gemini 3.8 Flash low effort 80.7% Grok Grok 4.6 low effort 80.3% Claude Claude Opus 5 xhigh effort 80.3% DeepSeek DeepSeek V4 Pro max effort 80.3% OpenAI GPT-5.6 Sol medium effort 80.3% OpenAI GPT-5.6 Luna high effort 80.3% Qwen Qwen3.8 2.4T A95B xhigh effort 80.3% Grok Grok 4.6 high effort 80.0% Gemini Gemini 3.6 Flash high effort 80.0% Zhipu GLM-5.3 Flash max effort 79.7% DeepSeek DeepSeek V4 Flash max effort 79.7% Qwen Qwen3.8 27B medium effort 79.7% XiaomiMiMo MiMo V2.5 Pro enabled effort 79.7% Zhipu GLM-5.3 max effort 79.3% Claude Claude Opus 5 max effort 79.3% Kimi Kimi K3 low effort 79.3% Kimi Kimi K2.7 Code enabled effort 79.3% Nvidia NVIDIA Nemotron 3 Ultra enabled effort 79.3% Grok Grok 4.5 high effort 79.0% Claude Claude Opus 5 high effort 79.0% Meta Muse Spark 1.2 xhigh effort 79.0% OpenAI GPT-5.6 Terra xhigh effort 79.0% Qwen Qwen3.7 Max unspecified effort 79.0% Tencent Tencent Hy3 high effort 78.7% Gemini Gemini 3.7 Flash low effort 78.3% Minimax MiniMax M2.7 Reasoning enabled 78.3% Qwen Qwen3.8 Max xhigh effort 78.3% Zhipu GLM-5.2 max effort 78.0% Kimi Kimi K2.5 enabled effort 78.0% OpenAI GPT-5.6 Sol low effort 77.7% Claude Claude Opus 4.8 max effort 77.7% OpenAI GPT-5.6 Terra high effort 77.3% Qwen Qwen3.6 27B enabled effort 77.3% Qwen Qwen3.8 27B low effort 77.3% TM Inkling max effort 76.7% OpenAI GPT-5.4 low effort 75.0% OpenAI GPT-5.6 Luna medium effort 75.0% Grok Grok 4.3 medium effort 74.0% OpenAI GPT-5.6 Terra medium effort 74.0% Grok Grok 4.3 low effort 73.7% StepFun Step 3.7 Flash high effort 73.7% Zhipu GLM-5.1 enabled effort 73.3% DeepSeek DeepSeek V3.2 enabled effort 73.3% Minimax MiniMax M2.5 Reasoning enabled 73.0% Grok Grok 4.3 high effort 73.0% XiaomiMiMo MiMo V2.5 enabled effort 72.0% DeepSeek DeepSeek V4 Flash high effort 71.7% Qwen Qwen3.6 35B A3B enabled effort 71.3% OpenAI GPT-5.6 Terra low effort 70.7% XiaomiMiMo MiMo V2 Flash enabled effort 70.0% OpenAI GPT-5.6 Luna low effort 69.7% Gemini Gemma 4 31B IT enabled effort 69.7% Kimi Kimi K2.6 disabled effort 69.3% Mistral Mistral Medium 3.5 high effort 67.3% Kimi Kimi K2.5 disabled effort 66.7% Qwen Qwen3.6 27B disabled effort 65.7% Nvidia NVIDIA Nemotron 3 Super enabled effort 64.3% Qwen Qwen3.6 35B A3B disabled effort 64.0% OpenAI GPT-5.5 none effort 62.3% OpenAI GPT-5.6 Sol none effort 60.3% Nova Amazon Nova 2 Lite High run 60.0% Nova Amazon Nova 2 Lite Medium run 58.7% OpenAI GPT-5.6 Terra none effort 58.3% OpenAI GPT-5.4 none effort 57.7% DeepSeek DeepSeek R1 Published default 54.3% Nova Amazon Nova 2 Lite Low run 53.3% Zhipu GLM-5.1 disabled effort 52.7% Cohere Command A+ 05-2026 Reasoning enabled 52.0% OpenAI GPT-OSS 120B high effort 50.0% Meta Llama 4 Maverick 17B 128E Instruct Non-reasoning run 50.0% StepFun Step 3.5 Flash Reasoning enabled 49.7% Mistral Mistral Small 4 high effort 46.7% Gemini Gemma 4 31B IT disabled effort 46.0% OpenAI GPT-OSS 120B low effort 45.7% DeepSeek DeepSeek V3.2 disabled effort 42.7% OpenAI GPT-5.6 Luna none effort 42.3% Zhipu GLM-5.2 disabled effort 41.7% XiaomiMiMo MiMo V2.5 Pro disabled effort 39.7% Nvidia Nemotron 3 Nano Omni 30B A3B Reasoning enabled effort 37.3% Cohere North Mini Code 1.0 Reasoning enabled 36.3% XiaomiMiMo MiMo V2 Flash disabled effort 36.0% Mistral Mistral Large 3 Published default 34.7% OpenAI GPT-OSS 20B high effort 34.3% Nova Amazon Nova Premier Non-reasoning run 32.3% Grok Grok 4.3 none effort 31.0% OpenAI GPT-OSS 20B low effort 28.3% Mistral Mistral Small 4 none effort 27.7% Meta Llama 4 Scout 17B 16E Instruct Non-reasoning run 21.0% Nova Amazon Nova Pro Published default 19.3% AI21 Jamba Large 1.7 Non-reasoning run 18.7% Nova Amazon Nova 2 Lite No reasoning 6.3% AI21 Jamba Reasoning 3B Reasoning enabled 0.0% Cohere Tiny Aya Global Non-reasoning run 0.0% StepFun Step3-VL 10B Reasoning enabled elo · Higher is better 108 results
63.5% Claude Claude Fable 5.1 max effort 62.6% Claude Claude Fable 5.1 xhigh effort 62.0% Claude Claude Opus 5 max effort 61.0% Meta Muse Spark 1.3 max effort 60.7% Claude Claude Opus 5 xhigh effort 59.0% Zhipu GLM-5.3 max effort 58.7% Zhipu GLM-5.3 Flash max effort 58.4% Grok Grok 4.6 xhigh effort 58.3% Meta Muse Spark 1.3 xhigh effort 57.8% Claude Claude Fable 5.1 high effort 57.4% Grok Grok 4.6 high effort 57.4% Grok Grok 4.6 medium effort 56.8% Claude Claude Fable 5 max effort 56.7% Qwen Qwen3.8 Max xhigh effort 56.7% Claude Claude Opus 5 high effort 56.6% Qwen Qwen3.8 2.4T A95B xhigh effort 56.4% OpenAI GPT-5.6 Sol max effort 54.5% OpenAI GPT-5.6 Sol xhigh effort 54.4% Kimi Kimi K3 max effort 54.3% Claude Claude Fable 5.1 medium effort 51.5% Claude Claude Opus 5 medium effort 51.4% OpenAI GPT-5.6 Sol high effort 51.4% Meta Muse Spark 1.2 xhigh effort 50.3% Claude Claude Fable 5.1 low effort 50.3% Claude Claude Sonnet 5 max effort 49.9% DeepSeek DeepSeek V4 Pro max effort 49.7% Claude Claude Opus 4.8 max effort 49.7% OpenAI GPT-5.6 Luna max effort 49.2% OpenAI GPT-5.6 Terra xhigh effort 49.1% OpenAI GPT-5.6 Terra max effort 48.6% DeepSeek DeepSeek V4 Flash max effort 48.4% Qwen Qwen3.8 27B xhigh effort 48.4% Grok Grok 4.6 low effort 48.3% Gemini Gemini 3.8 Flash high effort 48.0% OpenAI GPT-5.6 Sol medium effort 47.9% Gemini Gemini 3.8 Flash medium effort 46.8% Gemini Gemini 3.7 Flash high effort 46.7% Grok Grok 4.5 high effort 46.7% Qwen Qwen3.8 27B medium effort 46.7% OpenAI GPT-5.6 Luna xhigh effort 46.1% Zhipu GLM-5.2 max effort 46.0% OpenAI GPT-5.6 Terra high effort 45.7% Claude Claude Sonnet 5 xhigh effort 45.6% Gemini Gemini 3.7 Flash medium effort 45.4% Qwen Qwen3.8 27B low effort 45.0% OpenAI GPT-5.5 xhigh effort 44.0% OpenAI GPT-5.6 Luna high effort 43.8% Claude Claude Opus 5 low effort 43.8% OpenAI GPT-5.5 high effort 43.2% Gemini Gemini 3.8 Flash low effort 43.0% OpenAI GPT-5.6 Sol low effort 43.0% Gemini Gemini 3.7 Flash low effort 41.8% Gemini Gemini 3.6 Flash high effort 41.2% OpenAI GPT-5.6 Terra medium effort 41.0% Claude Claude Sonnet 5 high effort 40.6% Zhipu GLM-5.2 disabled effort 40.5% OpenAI GPT-5.4 xhigh effort 40.4% Minimax MiniMax M3 adaptive effort 40.2% OpenAI GPT-5.6 Sol none effort 39.7% OpenAI GPT-5.5 medium effort 36.3% Claude Claude Sonnet 5 medium effort 35.1% Kimi Kimi K3 low effort 35.0% OpenAI GPT-5.6 Luna medium effort 34.7% Qwen Qwen3.7 Max unspecified effort 34.5% XiaomiMiMo MiMo V2.5 Pro enabled effort 34.3% Zhipu GLM-5.1 enabled effort 34.2% OpenAI GPT-5.6 Terra low effort 33.7% OpenAI GPT-5.6 Terra none effort 33.4% TM Inkling max effort 32.5% Claude Claude Sonnet 5 low effort 32.0% Tencent Tencent Hy3 high effort 31.0% Kimi Kimi K2.6 enabled effort 30.9% Kimi Kimi K2.7 Code enabled effort 30.9% OpenAI GPT-5.5 low effort 29.8% Nvidia NVIDIA Nemotron 3 Ultra enabled effort 29.6% Minimax MiniMax M2.7 Reasoning enabled 29.4% OpenAI GPT-5.6 Luna low effort 29.3% DeepSeek DeepSeek V4 Flash high effort 29.2% XiaomiMiMo MiMo V2.5 enabled effort 28.6% Qwen Qwen3.6 27B enabled effort 27.9% OpenAI GPT-5.5 none effort 27.3% Qwen Qwen3.6 27B disabled effort 26.8% Grok Grok 4.3 none effort 26.1% Grok Grok 4.3 high effort 25.6% OpenAI GPT-5.6 Luna none effort 24.8% Qwen Qwen3.6 35B A3B enabled effort 22.9% StepFun Step 3.7 Flash high effort 22.9% Qwen Qwen3.6 35B A3B disabled effort 22.1% Kimi Kimi K2.5 enabled effort 19.8% Meta Muse Glimmer 30B high effort 18.9% Mistral Mistral Medium 3.5 high effort 15.7% DeepSeek DeepSeek V3.2 enabled effort 14.6% XiaomiMiMo MiMo V2 Flash disabled effort 12.9% Gemini Gemma 4 31B IT enabled effort 12.4% OpenAI GPT-OSS 120B high effort 9.7% Gemini Gemma 4 31B IT disabled effort 8.1% Cohere Command A+ 05-2026 Reasoning enabled 7.4% Nvidia NVIDIA Nemotron 3 Super enabled effort 4.6% Mistral Mistral Large 3 Published default 1.9% Mistral Mistral Small 4 high effort 1.9% Nova Amazon Nova 2 Lite High run 0.8% OpenAI GPT-OSS 20B high effort 0.0% Cohere North Mini Code 1.0 Reasoning enabled 0.0% DeepSeek DeepSeek R1 Published default 0.0% Nvidia Nemotron 3 Nano Omni 30B A3B Reasoning enabled effort 0.0% OpenAI GPT-OSS 120B low effort 0.0% Meta Llama 4 Scout 17B 16E Instruct Non-reasoning run 0.0% Meta Llama 4 Maverick 17B 128E Instruct Non-reasoning run resolved · Higher is better 104 results
91.4% Claude Claude Fable 5.1 max effort 91.0% Claude Claude Fable 5.1 xhigh effort 89.9% Claude Claude Fable 5.1 high effort 89.5% OpenAI GPT-5.6 Sol xhigh effort 89.1% Claude Claude Opus 5 max effort 88.4% Grok Grok 4.6 high effort 88.0% Claude Claude Fable 5.1 medium effort 88.0% Claude Claude Opus 5 xhigh effort 88.0% OpenAI GPT-5.6 Sol max effort 88.0% OpenAI GPT-5.6 Terra max effort 88.0% Grok Grok 4.6 xhigh effort 87.6% Claude Claude Opus 5 high effort 87.6% Gemini Gemini 3.8 Flash high effort 87.3% OpenAI GPT-5.6 Sol high effort 86.1% Claude Claude Opus 5 medium effort 86.1% OpenAI GPT-5.6 Sol medium effort 85.8% Gemini Gemini 3.7 Flash high effort 85.8% Meta Muse Spark 1.3 max effort 85.4% Meta Muse Spark 1.3 xhigh effort 85.0% Claude Claude Fable 5.1 low effort 85.0% Kimi Kimi K3 max effort 84.6% Claude Claude Fable 5 max effort 84.6% Claude Claude Opus 4.8 max effort 84.3% OpenAI GPT-5.5 xhigh effort 84.3% Grok Grok 4.6 medium effort 84.3% Zhipu GLM-5.3 Flash max effort 83.9% Gemini Gemini 3.8 Flash medium effort 83.9% Zhipu GLM-5.3 max effort 83.1% Gemini Gemini 3.8 Flash low effort 82.4% Kimi Kimi K3 low effort 82.0% Qwen Qwen3.8 2.4T A95B xhigh effort 81.6% Grok Grok 4.5 high effort 81.3% Qwen Qwen3.8 Max xhigh effort 80.9% OpenAI GPT-5.6 Luna max effort 80.5% Claude Claude Sonnet 5 max effort 80.5% OpenAI GPT-5.5 medium effort 80.1% Meta Muse Spark 1.2 xhigh effort 80.1% OpenAI GPT-5.6 Terra xhigh effort 79.8% Gemini Gemini 3.7 Flash low effort 79.8% Qwen Qwen3.8 27B xhigh effort 79.4% OpenAI GPT-5.5 high effort 78.7% DeepSeek DeepSeek V4 Pro max effort 78.7% DeepSeek DeepSeek V4 Flash max effort 78.3% Gemini Gemini 3.7 Flash medium effort 78.3% OpenAI GPT-5.4 xhigh effort 77.9% OpenAI GPT-5.6 Luna xhigh effort 77.9% Zhipu GLM-5.2 max effort 77.5% Gemini Gemini 3.6 Flash high effort 76.8% OpenAI GPT-5.6 Sol low effort 76.4% Claude Claude Opus 5 low effort 75.7% OpenAI GPT-5.6 Terra high effort 75.3% Grok Grok 4.6 low effort 74.5% Qwen Qwen3.7 Max unspecified effort 74.2% OpenAI GPT-5.6 Sol none effort 72.3% OpenAI GPT-5.6 Terra medium effort 69.7% OpenAI GPT-5.6 Luna high effort 67.4% Kimi Kimi K2.7 Code enabled effort 67.4% Qwen Qwen3.8 27B low effort 65.9% Kimi Kimi K2.6 enabled effort 65.5% OpenAI GPT-5.5 low effort 65.2% Minimax MiniMax M3 adaptive effort 65.2% Qwen Qwen3.8 27B medium effort 65.2% XiaomiMiMo MiMo V2.5 Pro enabled effort 64.4% Tencent Tencent Hy3 high effort 63.7% XiaomiMiMo MiMo V2.5 enabled effort 62.5% OpenAI GPT-5.6 Terra low effort 61.8% XiaomiMiMo MiMo V2 Flash disabled effort 61.8% Zhipu GLM-5.1 enabled effort 61.0% OpenAI GPT-5.5 none effort 60.7% Qwen Qwen3.6 27B enabled effort 56.9% DeepSeek DeepSeek V4 Flash high effort 56.2% OpenAI GPT-5.6 Terra none effort 55.4% Minimax MiniMax M2.7 Reasoning enabled 55.1% TM Inkling max effort 53.9% Nvidia NVIDIA Nemotron 3 Ultra enabled effort 53.2% OpenAI GPT-5.6 Luna medium effort 51.7% Meta Muse Glimmer 30B high effort 51.7% Zhipu GLM-5.2 disabled effort 51.3% Qwen Qwen3.6 27B disabled effort 50.6% Mistral Mistral Medium 3.5 high effort 46.8% DeepSeek DeepSeek V3.2 enabled effort 45.7% Kimi Kimi K2.5 enabled effort 44.9% Qwen Qwen3.6 35B A3B enabled effort 43.4% Gemini Gemma 4 31B IT enabled effort 43.4% OpenAI GPT-5.6 Luna low effort 41.6% Qwen Qwen3.6 35B A3B disabled effort 39.7% Grok Grok 4.3 high effort 39.3% StepFun Step 3.7 Flash high effort 39.0% OpenAI GPT-5.6 Luna none effort 38.6% Nvidia NVIDIA Nemotron 3 Super enabled effort 35.6% Cohere North Mini Code 1.0 Reasoning enabled 34.1% Grok Grok 4.3 none effort 29.2% Gemini Gemma 4 31B IT disabled effort 26.2% OpenAI GPT-OSS 120B high effort 22.8% Cohere Command A+ 05-2026 Reasoning enabled 21.0% Mistral Mistral Small 4 high effort 19.1% DeepSeek DeepSeek R1 Published default 16.1% Nova Amazon Nova 2 Lite High run 13.9% OpenAI GPT-OSS 120B low effort 13.9% OpenAI GPT-OSS 20B high effort 12.0% Mistral Mistral Large 3 Published default 7.9% Meta Llama 4 Maverick 17B 128E Instruct Non-reasoning run 6.7% Nvidia Nemotron 3 Nano Omni 30B A3B Reasoning enabled effort 3.7% Meta Llama 4 Scout 17B 16E Instruct Non-reasoning run score · Higher is better 100 results
52.4% Meta Muse Spark 1.3 max effort 51.3% Qwen Qwen3.8 Max xhigh effort 50.7% Grok Grok 4.6 high effort 50.3% Zhipu GLM-5.3 max effort 49.1% Qwen Qwen3.8 2.4T A95B xhigh effort 48.0% Qwen Qwen3.8 27B xhigh effort 47.4% Qwen Qwen3.8 27B medium effort 47.2% Claude Claude Fable 5.1 max effort 47.2% Meta Muse Spark 1.3 xhigh effort 47.2% Zhipu GLM-5.3 Flash max effort 46.0% Kimi Kimi K3 max effort 45.8% Claude Claude Fable 5.1 xhigh effort 45.8% Gemini Gemini 3.8 Flash medium effort 44.9% Gemini Gemini 3.8 Flash high effort 44.7% Claude Claude Opus 5 high effort 44.3% OpenAI GPT-5.6 Sol max effort 44.3% Grok Grok 4.6 medium effort 43.3% Claude Claude Opus 5 xhigh effort 43.3% Grok Grok 4.6 xhigh effort 43.1% Claude Claude Fable 5.1 high effort 42.1% Claude Claude Opus 5 max effort 42.1% Grok Grok 4.5 high effort 41.6% Kimi Kimi K3 low effort 41.0% Claude Claude Fable 5.1 medium effort 40.2% OpenAI GPT-5.6 Terra max effort 39.6% DeepSeek DeepSeek V4 Pro max effort 39.6% OpenAI GPT-5.4 xhigh effort 39.4% DeepSeek DeepSeek V4 Flash max effort 39.0% Claude Claude Fable 5.1 low effort 39.0% OpenAI GPT-5.5 xhigh effort 38.6% Claude Claude Opus 5 medium effort 38.1% Claude Claude Fable 5 max effort 38.1% OpenAI GPT-5.6 Sol xhigh effort 38.1% Grok Grok 4.6 low effort 37.3% Claude Claude Sonnet 5 max effort 36.7% OpenAI GPT-5.6 Sol high effort 36.7% OpenAI GPT-5.5 high effort 36.5% OpenAI GPT-5.6 Sol medium effort 35.5% Gemini Gemini 3.7 Flash medium effort 34.8% Meta Muse Spark 1.2 xhigh effort 34.6% Zhipu GLM-5.2 max effort 34.2% Claude Claude Opus 4.8 max effort 33.2% Gemini Gemini 3.8 Flash low effort 32.8% Gemini Gemini 3.7 Flash high effort 32.2% Qwen Qwen3.8 27B low effort 31.1% OpenAI GPT-5.6 Luna max effort 30.3% Claude Claude Opus 5 low effort 29.9% Gemini Gemini 3.6 Flash high effort 29.9% OpenAI GPT-5.5 medium effort 29.7% OpenAI GPT-5.6 Terra xhigh effort 29.5% Gemini Gemini 3.7 Flash low effort 29.1% OpenAI GPT-5.6 Sol low effort 29.1% TM Inkling max effort 28.7% OpenAI GPT-5.6 Terra high effort 28.7% OpenAI GPT-5.6 Luna xhigh effort 26.2% DeepSeek DeepSeek V4 Flash high effort 25.6% OpenAI GPT-5.6 Terra medium effort 25.2% OpenAI GPT-5.6 Luna high effort 24.9% OpenAI GPT-5.5 low effort 23.5% Meta Muse Glimmer 30B high effort 23.3% Kimi Kimi K2.6 enabled effort 22.9% Tencent Tencent Hy3 high effort 20.2% Kimi Kimi K2.7 Code enabled effort 19.6% OpenAI GPT-5.6 Sol none effort 18.8% OpenAI GPT-5.6 Terra low effort 17.7% OpenAI GPT-5.6 Luna medium effort 16.7% Qwen Qwen3.6 27B enabled effort 16.7% Zhipu GLM-5.2 disabled effort 15.7% OpenAI GPT-5.6 Terra none effort 15.3% Minimax MiniMax M3 adaptive effort 15.1% Mistral Mistral Medium 3.5 high effort 14.8% Gemini Gemma 4 31B IT enabled effort 14.8% OpenAI GPT-5.5 none effort 14.2% Kimi Kimi K2.5 enabled effort 14.2% Nvidia NVIDIA Nemotron 3 Ultra enabled effort 13.6% Zhipu GLM-5.1 enabled effort 12.8% OpenAI GPT-5.6 Luna low effort 12.8% OpenAI GPT-OSS 120B high effort 12.4% Grok Grok 4.3 high effort 12.0% StepFun Step 3.7 Flash high effort 11.8% Qwen Qwen3.7 Max unspecified effort 10.3% Nvidia NVIDIA Nemotron 3 Super enabled effort 9.9% Minimax MiniMax M2.7 Reasoning enabled 9.9% XiaomiMiMo MiMo V2.5 Pro enabled effort 9.7% OpenAI GPT-5.6 Luna none effort 9.3% Qwen Qwen3.6 27B disabled effort 9.3% Qwen Qwen3.6 35B A3B enabled effort 8.9% Gemini Gemma 4 31B IT disabled effort 8.7% XiaomiMiMo MiMo V2.5 enabled effort 8.0% Grok Grok 4.3 none effort 7.0% OpenAI GPT-OSS 20B high effort 6.4% Cohere North Mini Code 1.0 Reasoning enabled 6.4% DeepSeek DeepSeek R1 Published default 6.0% Cohere Command A+ 05-2026 Reasoning enabled 5.8% Mistral Mistral Large 3 Published default 5.4% Qwen Qwen3.6 35B A3B disabled effort 4.9% Mistral Mistral Small 4 high effort 3.7% Meta Llama 4 Maverick 17B 128E Instruct Non-reasoning run 3.3% Meta Llama 4 Scout 17B 16E Instruct Non-reasoning run 2.9% OpenAI GPT-OSS 120B low effort score · Higher is better 59 results
63.1% Claude Claude Fable 5.1 max effort 61.0% Claude Claude Fable 5 max effort 60.9% Claude Claude Fable 5.1 xhigh effort 59.8% Gemini Gemini 3.7 Flash medium effort 59.7% Meta Muse Spark 1.3 xhigh effort 59.5% Kimi Kimi K3 max effort 59.0% Zhipu GLM-5.3 max effort 58.7% Claude Claude Fable 5.1 high effort 58.3% Meta Muse Spark 1.3 max effort 57.8% OpenAI GPT-5.6 Sol high effort 57.4% Meta Muse Spark 1.2 xhigh effort 57.4% OpenAI GPT-5.6 Sol medium effort 57.2% Gemini Gemini 3.7 Flash high effort 57.1% OpenAI GPT-5.6 Sol max effort 57.1% OpenAI GPT-5.6 Sol xhigh effort 56.7% Claude Claude Fable 5.1 low effort 56.6% Gemini Gemini 3.8 Flash high effort 56.5% Grok Grok 4.6 high effort 56.4% Claude Claude Fable 5.1 medium effort 56.4% Claude Claude Opus 5 max effort 56.4% OpenAI GPT-5.6 Sol low effort 55.9% Grok Grok 4.6 medium effort 55.7% Claude Claude Opus 5 xhigh effort 55.7% Gemini Gemini 3.7 Flash low effort 55.4% Claude Claude Opus 5 high effort 55.1% Gemini Gemini 3.8 Flash medium effort 55.0% Gemini Gemini 3.8 Flash low effort 55.0% OpenAI GPT-5.6 Terra max effort 55.0% Grok Grok 4.5 high effort 54.3% Claude Claude Sonnet 5 max effort 54.1% Qwen Qwen3.8 2.4T A95B xhigh effort 53.6% OpenAI GPT-5.6 Luna max effort 53.4% Gemini Gemini 3.6 Flash high effort 53.2% Qwen Qwen3.8 Max xhigh effort 53.0% Grok Grok 4.6 xhigh effort 52.7% Kimi Kimi K3 low effort 52.4% OpenAI GPT-5.6 Terra high effort 52.3% OpenAI GPT-5.6 Terra xhigh effort 51.6% Zhipu GLM-5.3 Flash max effort 51.5% Claude Claude Opus 5 medium effort 51.2% Zhipu GLM-5.2 max effort 51.0% DeepSeek DeepSeek V4 Pro max effort 50.6% XiaomiMiMo MiMo V2.5 Pro enabled effort 50.5% OpenAI GPT-5.6 Luna xhigh effort 50.3% DeepSeek DeepSeek V4 Flash max effort 49.4% Grok Grok 4.6 low effort 49.2% Claude Claude Opus 5 low effort 47.1% Minimax MiniMax M3 adaptive effort 47.0% TM Inkling max effort 46.6% Qwen Qwen3.8 27B xhigh effort 45.5% Gemini Gemma 4 31B IT enabled effort 44.9% Meta Muse Glimmer 30B high effort 40.3% Nvidia NVIDIA Nemotron 3 Ultra enabled effort 40.2% Mistral Mistral Medium 3.5 high effort 40.0% Qwen Qwen3.8 27B low effort 39.0% Qwen Qwen3.8 27B medium effort 38.5% Cohere Command A+ 05-2026 Reasoning enabled 36.6% Mistral Mistral Large 3 Published default 34.0% OpenAI GPT-OSS 120B high effort resolved · Higher is better 20 results
80.6% DeepSeek DeepSeek V4 Pro max effort 80.2% Minimax MiniMax M2.5 Published default 80.2% Kimi Kimi K2.6 enabled effort 80.2% TM Inkling Small max effort 79.0% DeepSeek DeepSeek V4 Flash max effort 78.9% XiaomiMiMo MiMo V2.5 Pro enabled effort 78.0% Tencent Tencent Hy3 high effort 77.6% TM Inkling max effort 77.2% Qwen Qwen3.6 27B unspecified effort 76.8% Kimi Kimi K2.5 disabled effort 76.0% Meta Muse Glimmer 30B high effort 73.5% Azure MAI-Thinking-1 Effort not reported 73.4% Qwen Qwen3.6 35B A3B unspecified effort 73.4% XiaomiMiMo MiMo V2 Flash enabled effort 69.5% Nvidia NVIDIA Nemotron 3 Ultra high effort 56.0% ByteDance Seed OSS 36B Instruct Published default 53.2% OpenAI GPT-OSS 20B medium effort 52.6% OpenAI GPT-OSS 120B medium effort 51.6% Nvidia NVIDIA Nemotron 3.5 Lightning enabled effort 50.2% Nvidia Nemotron Cascade 2 30B A3B unspecified effort elo · Higher is better 25 results
58.3% Claude Claude Fable 5.1 max effort 57.9% Claude Claude Fable 5.1 xhigh effort 57.5% Claude Claude Opus 5 max effort 56.5% Claude Claude Opus 5 xhigh effort 54.2% Claude Claude Fable 5.1 high effort 53.1% Claude Claude Opus 5 high effort 53.0% Meta Muse Spark 1.3 max effort 52.6% Grok Grok 4.6 xhigh effort 52.0% Grok Grok 4.6 high effort 51.7% Claude Claude Fable 5 max effort 51.0% Zhipu GLM-5.3 max effort 50.1% Kimi Kimi K3 max effort 49.3% Meta Muse Spark 1.3 xhigh effort 49.1% OpenAI GPT-5.6 Sol max effort 48.0% Zhipu GLM-5.3 Flash max effort 47.3% Qwen Qwen3.8 2.4T A95B xhigh effort 45.2% Qwen Qwen3.8 27B xhigh effort 42.2% OpenAI GPT-5.6 Luna max effort 41.7% OpenAI GPT-5.6 Terra max effort 38.6% DeepSeek DeepSeek V4 Pro max effort 35.3% Gemini Gemini 3.8 Flash high effort 30.1% Minimax MiniMax M3 adaptive effort 16.9% TM Inkling max effort 0.0% Meta Muse Glimmer 30B high effort 0.0% OpenAI GPT-OSS 120B high effort all pass · Higher is better 21 results
28.2% OpenAI GPT-5.6 Sol max effort 27.8% OpenAI GPT-5.6 Sol high effort 26.2% Claude Claude Fable 5.1 max effort 25.6% Meta Muse Spark 1.3 max effort 25.6% OpenAI GPT-5.6 Terra max effort 24.0% Claude Claude Fable 5 max effort 23.4% Meta Muse Spark 1.3 xhigh effort 23.0% OpenAI GPT-5.6 Luna max effort 21.6% Claude Claude Opus 5 max effort 19.6% Kimi Kimi K3 max effort 19.0% Gemini Gemini 3.8 Flash high effort 18.8% Grok Grok 4.6 high effort 16.6% Qwen Qwen3.8 27B xhigh effort 15.0% Qwen Qwen3.8 2.4T A95B xhigh effort 13.4% Meta Muse Glimmer 30B high effort 13.0% DeepSeek DeepSeek V4 Pro max effort 13.0% Zhipu GLM-5.3 Flash max effort 12.6% TM Inkling max effort 11.8% Zhipu GLM-5.3 max effort 10.4% Minimax MiniMax M3 adaptive effort 3.6% OpenAI GPT-OSS 120B high effort score · Higher is better 24 results
83.3% Claude Claude Fable 5 xhigh effort 66.6% Claude Claude Opus 4.8 xhigh effort 66.1% Claude Claude Fable 5 medium effort 57.2% Grok Grok 4.5 medium effort 55.8% Gemini Gemini 3.6 Flash medium effort 52.8% Meta Muse Spark 1.2 medium effort 47.6% Kimi Kimi K3 high effort 47.4% Claude Claude Opus 4.8 medium effort 46.4% Grok Grok 4.5 low effort 42.7% Qwen Qwen3.7 Max enabled effort 42.1% Qwen Qwen3.7 Max disabled effort 40.9% Claude Claude Sonnet 5 medium effort 40.2% Meta Muse Spark 1.2 minimal effort 40.1% Qwen Qwen3.8 Max medium effort 37.8% Gemini Gemini 3.6 Flash minimal effort 37.6% Minimax MiniMax M3 disabled effort 34.0% TM Inkling none effort 32.3% StepFun Step 3.7 Flash medium effort 31.9% TM Inkling Small none effort 31.9% TM Inkling Small medium effort 31.0% StepFun Step 3.7 Flash low effort 30.7% Mistral Mistral Medium 3.5 none effort 28.0% Minimax MiniMax M3 enabled effort 24.7% TM Inkling medium effort accuracy · Higher is better 14 results
87.5% DeepSeek DeepSeek V4 Pro max effort 87.1% Kimi Kimi K2.5 enabled effort 86.2% DeepSeek DeepSeek V4 Flash max effort 86.2% Qwen Qwen3.6 27B unspecified effort 85.2% Gemini Gemma 4 31B IT enabled effort 85.2% Qwen Qwen3.6 35B A3B unspecified effort 85.0% Azure MAI-Thinking-1 Effort not reported 84.9% XiaomiMiMo MiMo V2 Flash enabled effort 84.3% Baidu ERNIE 5.1 Effort not reported 83.8% Baidu ERNIE 5.0 Effort not reported 82.7% ByteDance Seed OSS 36B Instruct Published default 81.9% Nvidia NVIDIA Nemotron 3.5 Lightning enabled effort 79.8% Nvidia Nemotron Cascade 2 30B A3B unspecified effort 67.2% Tencent Hunyuan A13B Instruct unspecified effort resolved · Higher is better 10 results
65.7% Tencent Tencent Hy4 Preview high effort 61.7% Qwen Qwen3.8 27B xhigh effort 57.9% Tencent Tencent Hy3 high effort 57.2% XiaomiMiMo MiMo V2.5 Pro enabled effort 56.1% XiaomiMiMo MiMo V2.5 enabled effort 55.9% TM Inkling Small max effort 54.3% TM Inkling max effort 52.8% Azure MAI-Thinking-1 Effort not reported 51.2% Meta Muse Glimmer 30B high effort 46.9% ByteDance Seed2.0 Pro Effort not reported