Step 3.7 Flash

StepFun's high-efficiency multimodal MoE model (196B language backbone, ~11B active) with native image and video understanding.

Select Provider