Models
DeepSeek V4 Pro 202606
Vendor-direct DeepSeek V4 Pro for deeper reasoning and code work.
deepseek-v4-pro-202606
Use modelDeepSeek V4 Flash 202605
Vendor-direct DeepSeek V4 Flash for quick, low-cost assistant work.
deepseek-v4-flash-202605
Use modelGemini 3.5 Flash
A quick general-purpose model for everyday coding assistant work.
gemini-3.5-flash
Use modelDeepSeek V4 Pro
Deep reasoning and structured output for heavier implementation work.
deepseek-v4-pro
Use modelDeepSeek V4 Flash
Low-cost fast responses for lightweight code edits and daily assistants.
deepseek-v4-flash
Use modelGLM 5.1
Long-form reasoning coverage for structured code and planning tasks.
glm-5.1
Use modelKimi K2.6
A strong coding model for code generation, repair, and review loops.
kimi-k2.6
Use modelKimi K2.5
Agentic coding and multimodal reasoning with a long working context.
kimi-k2.5
Use modelMiniMax M3
Large-context multimodal reasoning for long documents and agent plans.
minimax-m3
Use modelMiniMax M2.7
Balanced MiniMax model for long context reasoning and function calling.
minimax-m2.7
Use modelMiniMax M2.5
Compact MiniMax reasoning model for efficient agent and document work.
minimax-m2.5
Use modelGLM 5V Turbo
Vision-capable GLM 5 Turbo for multimodal analysis and tool use.
glm-5v-turbo
Use modelGLM 5 Turbo
Faster GLM 5 variant for everyday reasoning and structured responses.
glm-5-turbo
Use modelGLM 5
Flagship GLM reasoning model for long planning and tool-heavy work.
glm-5
Use modelDeepSeek V3.2
Deep reasoning and structured output for complex coding and planning.
deepseek-v3.2
Use model