← All model drops
Live drop
Step 5 Preview
StepFunDropped Sep 18, 2026
1M-context multimodal agent that ingests docs, video and code to ship long-horizon tasks end-to-end.
About this release
Live drop
StepFunDropped Sep 18, 2026
1M-context multimodal agent that ingests docs, video and code to ship long-horizon tasks end-to-end.
About this release
Step 5 Preview is StepFun’s flagship reasoning model for agentic work, with native text, image and video input and a 1M-token context window. It supports tool calling, structured output and long-document analysis for software engineering, finance and professional knowledge tasks. The model is served via the StepFun Open Platform API under the ID step-5-preview.
Read the official StepFun launch pageFirst builds incoming — this lane fills in automatically as people start building with Step 5 Preview. Know one? Submit the URL.