Step 5 Preview supports a 1 million token context window and vision capabilities to automate professional knowledge work and software engineering. The Mixture-of-Experts architecture employs 27 billion active parameters to reduce task costs while maintaining a score of 44 on the Artificial Analysis Intelligence Index. API access through StepFun AI Studio is currently the primary method of deployment before the scheduled weight release.

Early tests showed the model efficiently handles long-horizon coding tasks, including identifying root causes within 368,000 tokens of synthetic incident tickets in 90 seconds. Side-by-side comparisons with GLM 5.3 indicated a better ability to signal task completion without exceeding execution limits. Pricing for the API is $1.00 per 1 million input tokens and $2.70 per 1 million output tokens.