OpenAI-compatible API across GPU prefill and specialized decode silicon
General Compute's self-serve API exposes an OpenAI-compatible endpoint that routes inference across GPU prefill and specialized decode hardware such as SambaNova and Cerebras.
For people
For agents