Step 5 Preview – API Pricing & Suppliers

Step 3.7 Flash is StepEnjoyable’s newest high-efficiency multimodal Mixture-of-Experts mannequin. It pairs a 196B-parameter language spine with a imaginative and prescient encoder for native picture and video understanding, activating roughly 11B parameters per token. The mannequin helps a 256K context window and exposes selectable reasoning ranges (excessive/medium/low), letting callers commerce off pace, price, and depth of reasoning.

Designed for coding, agentic workflows, structured outputs, and long-context productiveness duties.