Description
Step 5 Preview is the flagship AI model developed by StepFun, specifically designed for agentic work. This model is particularly strong in software engineering, professional knowledge work, and finance. It features a sparse Mixture-of-Experts (MoE) architecture with 600 billion total parameters and 27 billion active parameters, allowing it to handle up to 1 million tokens of context. This enables the model to process text, images, and video inputs, making it highly versatile for complex tasks.
Read More
In software engineering, Step 5 Preview supports repository-level tasks, frontend and visual development, and programming for devices. It can navigate large codebases, coordinate changes across files, run and verify code, and iterate on results. The model's long-horizon agentic execution capabilities allow it to perform tasks that require repeated tool calls, execution, verification, and revision, drawing on long-context information throughout the workflow.
For professional knowledge work, Step 5 Preview combines research, multimodal understanding, and tool use to analyze long documents, charts, screenshots, images, and video, producing structured reports and analyses. Its focus on finance makes it particularly adept at tasks that combine quantitative reasoning with the analysis of financial data and supporting documents.
Step 5 Preview can be deployed through an OpenAI-compatible API using frameworks like vLLM or SGLang, or used for offline generation with the standalone Transformers implementation. It supports various deployment configurations, including FP8 and BF16 precision, across different hardware setups. The model is licensed under the StepFun Model License 1.0, which is an Apache-2.0-Style license with additional conditions specific to model use, modification, and distribution.
- CategoryAI Models & LLMs
- DeploymentCloud
- PlatformsWeb
Step 5 Preview Highlights
Sparse Mixture-of-Experts architecture
600B total parameters
27B active parameters
Supports up to 1M tokens of context
Multimodal input: text, image, video
Repository-level software engineering support
Long-horizon agentic execution
Strong in finance and professional knowledge work
Getting Started with Step 5 Preview
Access model: Visit Hugging Face
Authenticate: Sign in with credentials
Set up environment: Configure hardware and software
Integrate via API: Use OpenAI-compatible API
Optimize: Adjust inference settings for performance
Step 5 Preview's Use Cases
- Software Engineering
- Professional Knowledge Work
- Finance
- Long-Horizon Execution
- Multimodal Processing
