rank,name,org,type,access,overall,categorization,vendor,context,coa,splits,tax,reliability,latency_s,cost_index 1,Jupid Agent,Jupid,Agent,Hosted agent,94.6,95.4,94,96.2,94.7,93,94,92,4.5,7.6 2,Claude Mythos 5.1,Anthropic,Model,Partner access,94.4,94.9,93.9,94,94.1,94.6,93.8,90,6.2,10.9 3,Claude Fable 5.1,Anthropic,Model,Public API,94.2,94.7,93.8,93.8,93.9,94.4,93.6,90,5.9,10.6 4,GPT-6 Astra Preview,OpenAI,Model,Limited preview,94.1,94.5,93.7,93.4,93.7,95,93.4,88,6.6,13.5 5,GPT-5.6 Sol Pro,OpenAI,Model,Public API,93.8,94.2,93.6,93,93.4,94.5,93.1,88,7,14.6 6,Claude Opus 5,Anthropic,Model,Public API,93.6,94,93.5,93.2,93.3,93.8,93,89,5.4,9.4 7,GPT-5.6 Sol,OpenAI,Model,Public API,93.3,93.7,93.4,92.4,92.9,94,92.5,87,5.8,12 8,Grok 4.6,xAI,Model,Public API,92.9,93.4,93.3,91.9,92.5,93.4,92.1,85,4.9,7.4 9,Claude Sonnet 5,Anthropic,Model,Public API,92.5,93,93.1,91.8,92.1,92.8,91.7,86,3.6,4.6 10,Claude Opus 4.8,Anthropic,Model,Public API,92.2,92.6,93,91.2,91.6,92.4,91.2,85,5.2,9.5 11,DeepSeek V4-Pro,DeepSeek,Model,Open weights · vendor API,92,92.5,92.9,90.8,91.5,92.3,91,84,4.8,6.2 12,Gemini 3.1 Pro Preview,Google,Model,Public preview,91.6,92,92.5,90.4,91.2,92,90.7,83,5,8.5 13,Gemini 3.8 Flash,Google,Model,Public API,91.4,91.8,92.4,90.2,90.9,91.8,90.5,83,2.6,3.2 14,Qwen3.8-Max,Alibaba,Model,Vendor API,91,91.5,92.1,89.8,90.6,91.3,90.1,82,4.3,5.8 15,Kimi K3,Moonshot AI,Model,Open weights · vendor API,90.6,91,91.7,89.4,90.2,90.9,89.7,81,4.4,5.4 16,Grok 4.5,xAI,Model,Public API,90.3,90.8,91.4,89,89.9,90.7,89.4,80,4.6,6.4 17,GPT-5.6 Terra,OpenAI,Model,Public API,90,90.4,91,88.7,89.5,90.6,89.1,80,3.8,6.6 18,Gemini 3.7 Flash,Google,Model,Public API,89.6,90,90.7,88.3,89.1,89.9,88.6,79,2.5,3 19,Muse Spark,Meta,Model,Public API,89.2,89.6,90.3,87.9,88.7,89.5,88.2,78,3.9,5 20,Qwen3.8-Flash,Alibaba,Model,Open weights · vendor API,88.8,89.2,89.9,87.5,88.3,89.1,87.8,77,3,3.4 21,DeepSeek V4-Flash,DeepSeek,Model,Open weights · vendor API,88.4,88.9,89.6,87,87.9,88.7,87.4,76,3,3.3 22,GPT-5.6 Luna,OpenAI,Model,Public API,88,88.4,89.1,86.6,87.5,88.4,87,76,1.8,2.2 23,Mistral Medium 3.5,Mistral,Model,Public API,86.2,86.6,87.3,84.8,85.7,86.5,85.2,73,3.5,4.2 24,Gemini 3.5 Flash-Lite,Google,Model,Public API,85.8,86.2,86.9,84.4,85.3,86.1,84.8,72,1.6,1.7 25,Mistral Large 3,Mistral,Model,Open weights · vendor API,85.3,85.7,86.4,83.9,84.8,85.6,84.3,71,4,4.6 26,Cohere Command A+,Cohere,Model,Open weights · vendor API,84.5,84.9,85.6,83.1,84,84.8,83.5,70,3.6,4 27,Amazon Nova 2 Pro,Amazon,Model,Public API,84,84.4,85,82.6,83.5,84.3,83,69,3.3,3.6 28,Claude Haiku 4.5,Anthropic,Model,Public API,83.4,83.8,84.5,82,82.9,83.7,82.4,68,2.1,2.5 29,Perplexity Sonar Reasoning Pro,Perplexity,Model,Public API,83,83.4,84,81.6,82.5,83.2,82,67,4.4,4 30,Mistral Small 4,Mistral,Model,Open weights · vendor API,82.6,83,83.6,81.2,82.1,82.9,81.7,66,2.6,2.7