The Model Catalog.
One Unified API.

DeepSeek's sparse mixture-of-experts model activates 13B of its 284B parameters per request. This re-post-trained GA release is designed for coding, reasoning, and agent workflows, with a 1.31M-token context window.

GLM-5.3 is Z.ai's large-scale reasoning model for complex software engineering and long-horizon agent tasks. It supports text input and output with a 1M-token context window, while improving coding performance and the balance between capability and token efficiency over the previous generation. Reasoning is always enabled, with low, high, and max effort levels.

GLM-5.3-Flash is Z.ai's compute-efficient native multimodal model for coding and long-horizon agent workflows. Its hybrid sparse and linear attention architecture preserves long-context accuracy while reducing inference overhead across a 1M-token context window.

Moonshot AI's next-generation multimodal model for long-horizon coding, coding-driven UI/UX generation, and multi-agent orchestration. It handles complex end-to-end development across Python, Rust, and Go.

MiniMax's multimodal foundation model accepts text, images, and video and returns text across a 1M-token context window. It is suited for long-horizon agentic work, coding, tool use, and sustained multi-turn collaboration.

Qwen's open-weight dense vision-language model is designed for coding, professional workflows, research, multimodal interaction, and long-running agents. Supported clients can flexibly enable or disable thinking.