Step 5 Preview
Step 5 Preview is a large language model made by StepFun, added to the catalog on 8 Oct 2026. It has a context window of 1,000,000 tokens and costs $1.00 per 1M input tokens and $2.70 per 1M output tokens. It accepts text, images and video and produces text. Typical uses: reviewing and working with software code, interpreting images alongside text instructions and analyzing video with text prompts.
Price checked ยท source: OpenRouter
What Step 5 Preview is
StepFun: Step 5 Preview is a multimodal model from StepFun, described by its provider as a flagship model for agentic work. It accepts text, images, and video as input and produces text. Its stated areas of strength include software engineering and professional work; the available description does not specify further professional domains, so its fit is clearest for tasks involving code and mixed-media inputs.
It has a context window of 1,000,000 tokens and can generate up to 64,000 tokens in a response. This gives it room to work with lengthy prompts or collections of supplied material, while the output limit supports detailed text responses. The context figure describes how much input context it can hold, not a guarantee about how a particular long task will be handled.
Step 5 Preview supports reasoning, tool use, structured outputs, and response-format controls. These options suit workflows that need a model to work with tools or return text in a specified structure. Image and video inputs can accompany text-based requests, with text as the listed output modality. The provider describes the model as built on a sparse Mixture-of-Experts architecture, with 27 billion active parameters and 600 billion total parameters. These details identify the model but do not specify how a particular application should configure it. Potential uses include reviewing code, interpreting images or video alongside instructions, processing long documents, and producing structured text for downstream workflows.
Key features
- Accepts text, image, and video inputs and produces text.
- Has a 1,000,000-token context window.
- Supports up to 64,000 output tokens.
- Supports reasoning, tools, structured outputs, and response-format controls.
- Uses a sparse Mixture-of-Experts architecture with 27B active and 600B total parameters.
Typical uses
- Reviewing and working with software code
- Interpreting images alongside text instructions
- Analyzing video with text prompts
- Processing long documents
- Generating structured text for workflows
Step 5 Preview pricing
US dollars, list price through OpenRouter.
| Item | Standard |
|---|---|
| Input, per 1M tokens | $1.00 |
| Output, per 1M tokens | $2.70 |
| Cached input read, per 1M tokens | $0.10 |
No price change since it was added on 8 Oct 2026.
| What a task costs | Cost |
|---|---|
| Answer a short question500 in / 300 out | $0.0013 |
| Summarise a 50-page PDF25,000 in / 800 out | $0.027 |
| Review a 2,000-line source file30,000 in / 2,000 out | $0.035 |
| Read a 300-page contract set150,000 in / 1,500 out | $0.154 |
Step 5 Preview specifications
| Model id | stepfun/step-5-preview |
|---|---|
| Version | step-5-preview-20261008 |
| Context window | 1,000,000 tokensAbout 750,000 English words |
| Max output | 64,000 tokens |
| Knowledge cutoff | Not published |
| Added to catalog | 8 Oct 2026 |
| Input types | text, image, video |
|---|---|
| Output types | text |
| Tool calling | Yes |
| Structured outputs | Yes |
| Reasoning effort control | Yes |
| Content moderation | Not applied by provider |
Sample outputs
The same prompts are given to every model, and the unedited output is shown with the date it was produced. Upend does not publish leaderboard scores it did not run.
No output has been recorded for Step 5 Preview yet.
Compare Step 5 Preview
Side by side with other models: prices, context window, what each accepts and supports, and the same four tasks priced on both.