StepFun
A Chinese AI lab making open multimodal and reasoning models.
- Type
- AI lab
- Last shipped
- Oct 15, 2026
Everything StepFun shipped
October 2026
- Availability
StepFun released Step 5 Preview as open weights, with BF16 and FP8 versions on Hugging Face and ModelScope.
- Availability
StepFun made Step 5 Preview available on OpenRouter, with a week of free access rolling out across several coding tools.
September 2026
- Launch
StepFun open-sourced onPanda, its internal tool for annotating LLM training data and inspecting model outputs token by token.
- Launch
StepFun released Step Code v0.1.0, an open-source command line tool under the MIT License that covers coding, testing and publishing static sites.
- New model
StepFun introduced Step 5 Preview, a new flagship model for agentic work with 600B total and 27B active parameters, 1M context and vision.
- New model
StepFun and ACE Studio released StepAudio 3 Music, StepFun's first music generation model, which turns a prompt and lyrics into a complete song.
- New model
StepFun launched StepAudio 3, a family of 5 audio models for real-time voice, speech recognition, speech generation, audio generation and music, available now.
