Model wiki,know before you choose

OpenAI's new flagship general model, balanced in reasoning, creation and tool use

Premium tier of GPT-5 with stronger reasoning and stability for hard tasks

Lightweight GPT-5 with low latency and high cost-efficiency for high-frequency calls

Smallest GPT-5 tier for simple tasks and edge scenarios

Software-engineering agent model deeply integrated into coding workflows

OpenAI's open-weight model, deployable on self-hosted infrastructure

Premium tier of OpenAI's reasoning family with deep slow-thinking

Anthropic's balanced flagship, strong at writing and long context

Anthropic's most powerful flagship, ceiling for complex tasks

Lightweight fast model with quick responses and low price

Previous balanced flagship, stable value pick

Previous top-tier model for demanding tasks

Lightweight tier of the Claude 3 family, classic low-cost choice

Google's flagship model, balanced multimodality and long context

Fast, low-cost multimodal model for high-frequency scenarios

Lightest Gemini tier for large-scale simple calls

Image-generation edition of Gemini 3, native multimodal creation

Gemini 3 fast-tier preview, low-latency multimodal

Open-source reasoning model driven by RL-based deep thinking

General open-source MoE model with high cost-efficiency

New-generation flagship on par with leading international models

Fast tier of the new generation, low latency and cost

Main chat model with highly competitive pricing

Alibaba's latest Qwen flagship with strongest overall capability

Flagship MoE model, strong at long text and multilingual tasks

Qwen flagship, balanced reasoning and creation

Mid-tier workhorse balancing capability and cost

Code-specialized model with excellent engineering performance

Vision-language model for image+text understanding and document parsing

Classic flagship family with strong Chinese understanding

Kimi's latest flagship with ultra-long context and deep thinking

Balanced flagship for long text and tool use

Open-source flagship strong at agentic tasks

Rolling model following Moonshot's latest release

Zhipu's latest flagship with upgraded reasoning and coding

New-generation flagship with balanced general capability

Flagship family with long-text and agentic capabilities

Main model balancing capability and cost

Previous workhorse, stable and reliable

Doubao's latest flagship with leading overall capability

Flagship model, balanced multimodality and reasoning

Main flagship with outstanding value

Code-specialized model with strong engineering capability

Vision model for image Q&A and OCR

MiniMax's latest flagship with long context and MoE architecture

Main model balancing performance and cost

xAI's latest flagship with real-time info and multimodality

Previous flagship with upgraded reasoning