The newest and strongest from every lab.
Full-modal Qwen at its strongest: text, image, audio, video.
Refined Fable generation for long-running agentic work.
Anthropic flagship for hard reasoning and long-horizon agents.
OpenAI flagship for demanding end-to-end work.
StepFun’s next-gen flagship base model: text, image, video in.
770B-A49B Hunyuan generation for agents and coding.
No per-vendor accounts. No per-vendor SDKs. One key opens every door.
Change two lines. That's it.
/v1/No tricks. No hidden fees. What you see is what you get.
Every response includes the actual model field. Verify you got what you paid for. We never substitute cheaper models.
Prompt content and response bodies pass through to the model and are discarded. We log only token counts for billing. Read policy
Token prices match each lab's published rates. No hidden markups on usage - what you see is what you pay.