StepFun has open-sourced OnPanda, a tool for LLM data annotation and model inspection that enables token-level correction and supervision. The tool reduces annotation time by 52% compared to manual editing and supports multimodal data annotation across image, audio, and video with precise token-level control and steering.
StepFun announced Step 5 Preview, a 600B parameter mixture-of-experts model with 27B active parameters, designed for agentic work and professional knowledge tasks including software engineering and finance. The model features 1M context window, vision capabilities, and will be released as open weights on October 15.
StepFun's Step 5 Preview is a proprietary reasoning model with 600B parameters released September 18, 2026, scoring 44 on the Artificial Analysis Intelligence Index with competitive pricing of $1.00 per 1M input tokens and $2.70 per 1M output tokens. The multimodal model supports text and image inputs, offers a 1M token context window, and ranks well above average in intelligence compared to similarly priced models.