Models

State-of-the-art reasoning model with chain-of-thought. Matches GPT-o1 at 95% lower cost.

DeepSeek's flagship general-purpose model. Exceptional coding and instruction following.

Alibaba's flagship MoE model. 235B total params, 22B active. Best-in-class multilingual performance.

Baidu's latest multimodal model. Superior Chinese language understanding and generation.

Next-gen agentic model for autonomous multi-step task execution. Excels at coding and planning.

Ultra-fast and ultra-cheap. Perfect for high-volume, simple tasks.

ByteDance's flagship Doubao Seed 2.0 Pro — strong reasoning and instruction following. Live endpoint.

Free tier of DeepSeek R1. Rate limited but great for testing and low-volume use cases.

Qwen3's dense 72B model. Strong reasoning and coding with excellent multilingual support.

Free, efficient 8B model from Alibaba. Great for lightweight tasks and high-throughput applications.

Ultra-fast, ultra-cheap model for high-volume simple tasks. Best price-performance ratio.

Zhipu's premium model. Strong reasoning and instruction following.

ByteDance's enterprise model with ultra-long context. Perfect for document analysis.

Lightweight, cost-efficient model from ByteDance. Great for simple chat and classification tasks.

SenseTime's latest flagship model with strong reasoning and long context capabilities.

iFlytek's most capable model. Strong at Chinese language tasks and multimodal understanding.

Moonshot's frontier reasoning model. Exceptional at math, science, and complex problem solving.

01.AI's fastest model. Optimized for speed and cost efficiency in production deployments.

Tencent's latest Hunyuan model with multimodal capabilities and long context support.