Submit

Chinese reasoning models

Models that expose an explicit `<think>` or chain-of-thought trace — DeepSeek-R1, Qwen-QwQ, GLM-Z1. Best for math-heavy, multi-step, or agentic workloads where the reasoning trace is itself valuable. Includes tool-use / function-calling tuned variants.

ModelCreatorContextOpen weightTags
Yi Lightning01.AI16K—chat · reasoning
Qwen VL MaxAlibaba Cloud131KYeschat · vision · ocr
Qwen 2.5 MaxAlibaba Cloud33K—chat · reasoning · code
Qwen3.5-122B-A10BAlibaba Cloud262KYeschat · vision · tool_calling
Qwen3.5 397B A17BAlibaba Cloud262KYeschat · reasoning · code
Qwen3 72B InstructAlibaba131KYeschat · reasoning · multilingual
Doubao Pro 32KByteDance33K—chat · reasoning · code
DeepSeek V3.1DeepSeek33KYeschat · reasoning · code
R1DeepSeek64KYesreasoning · math · coding
DeepSeek V3DeepSeek AI66KYesreasoning · coding · tool_use
DeepSeek V4 ProDeepSeek1049KYeschat · reasoning · code
Llama 3.1 405B InstructMeta131KYeschat · reasoning · code
Llama 3.3 70B InstructMeta131KYeschat · coding · reasoning
Kimi K2 0711Moonshot AI131KYestext generation · instruction following · agentic tasks
Kimi K2.5Moonshot AI262KYeschat · reasoning · code
Kimi K2.6Moonshot AI262KYeschat · reasoning · code
Step 3.5 FlashStepFun (阶跃星辰)262KYeschat · reasoning · code
GLM-4-AirZhipu AI131K—chat · reasoning · code
GLM 5Zhipu AI203KYesreasoning · coding · tool_calling
GLM 5.1Zhipu AI203KYeschat · reasoning · long_context

Other capabilities