Every model on the board.

Flagship reasoning model with visible chain of thought, strong at code and long-form work.

Moonshot's flagship open-weight model for long-context reasoning and agentic tool use.

Higher-throughput variant of Kimi K2 for latency-sensitive production traffic.

Vision-language variant of Kimi for document and image understanding.

Zhipu's general-purpose open-weight model, tuned for coding and long documents.

Lightweight GLM variant for high-volume, cost-sensitive workloads.

Vision-enabled GLM for chart, screenshot, and document reasoning.

Open-weight embedding model tuned for Chinese and English retrieval.

General-purpose mixture-of-experts model with strong coding performance.

Reasoning-focused model that shows its chain of thought before answering.

Alibaba's largest open-weight Qwen model for complex reasoning tasks.

Qwen variant fine-tuned specifically for repository-scale coding tasks.

Vision-language Qwen model for image, chart, and video-frame understanding.

Open-weight text-to-image generation model from the Qwen family.

Open-weight speech understanding and transcription model.

Open-weight multilingual embedding model for search and RAG.

Meta's flagship open-weight mixture-of-experts model.

Smaller, faster Llama 4 variant for everyday text tasks.

Mistral's top-tier open-weight model for reasoning and coding.

Sparse mixture-of-experts model balancing cost and capability.

Mistral's hosted embedding model for semantic search.

Open-weight long-context model tuned for agentic workflows.

StepFun's general-purpose open-weight language model.

Large open-weight model optimized for synthetic data and reasoning.

Baidu's open-weight ERNIE model for bilingual reasoning tasks.

ByteDance's open-weight Seed model for general text tasks.

ByteDance's text/image-to-video generation model.

ByteDance's high-fidelity text-to-image generation model.

ByteDance's text-to-speech model with multilingual voices.

Text and image-to-video generation model.

Most capable Claude model for complex reasoning and extended coding tasks.

Balanced Claude model for everyday production traffic.

Fast, cost-efficient Claude model for latency-sensitive assistants.

OpenAI's flagship model for advanced reasoning and coding.

Smaller, faster GPT-5.4 variant for high-volume traffic.

Open-weight speech-to-text model from OpenAI.

High-dimensional embedding model for search and clustering.

Google's flagship multimodal model with a 1M-token context window.

Faster, lower-cost Gemini variant for high-throughput tasks.

Google's hosted embedding model for retrieval and classification.