StepFun: Step 5 Preview

StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window

Compare models

step-5-preview

AvailableStepFunChatPreview

Input / Output: $1.00 / $2.70
Context: 1M
Max output: 64K
Modalities: VisionChat
Capabilities: Tool usePrompt CacheReasoning

About Step 5 Preview

Step 5 Preview is StepFun's flagship model for agentic work, aimed at software engineering and professional knowledge work, with particular strength in finance. It accepts text, images, and video, supports adjustable reasoning effort, and ships as a preview, so behavior may change before the final release. It sits above the Step 3.5 and 3.7 Flash models.

Where it works well

Native text, image, and video input: StepFun's documentation lists up to 60 images per request and video files under 128 MB.
Long-document analysis and cross-source research, where the very large context window carries whole reports.
Software troubleshooting and debugging across large codebases, with tool calling for multi-step tasks.
Reasoning effort can be set to low, medium, or high.
JSON mode, JSON Schema output, and prompt caching are supported.

When to choose another model

It is a preview release, so behavior can change between revisions; keep a fixed evaluation set and a fallback model.
StepFun says video understanding works best on clips under five minutes.
The model does not reach external services or your machine by itself; the application must supply those tools.

Getting started

Create API key
Create a key in Console, then use it with every model on the platform.

Send your first request

Copy the example for your language and run it against the endpoint.

Text to textMessages API

POST/v1/messages

API format:

curl https://api.tokenlab.sh/v1/messages \
  -H "Content-Type: application/json" \
  -H "x-api-key: sk-xxx" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "step-5-preview",
    "max_tokens": 1024,
    "messages": [
      {"role": "user", "content": "Hello!"}
    ]
  }'

Pricing

The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.

Default spec

per 1M tokens

Official price: Input $1.00 / Output $2.70 / Cache read $0.05
Verified price: —
Discount: —

Prompt cache pricing

Cache read

Official price: $0.05
Verified price: —
Discount: —

	Official priceper 1M tokens	Verified priceper 1M tokens	Discount
Default spec	Input $1.00 / Output $2.70 / Cache read $0.05	—	—
Prompt cache pricing
Cache read	$0.05	—	—

Usage & activity

Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.

Usage & availability

Last 24 hours

Requests
Success rate
P95 latency
Total tokens

Model performance

Last updated Nov 13, 2026, 4:51 AM UTC

Metrics appear once privacy and data volume thresholds are met.

Data is based on aggregate user requests, excluding status checks.

Open in Console

Open Step 5 Preview in Console with a prompt ready to edit or send.

Help me try step-5-preview with a short message at /v1/messages. Show the reply, latency, and cost.

Use cases

Best for

Vision
Reasoning

Financial document analysis
Feed annual reports, filings, and spreadsheets together and ask for reconciled figures and risks, using the finance focus StepFun highlights.
Codebase debugging
Load a large repository slice and a stack trace, then ask for the root cause and a narrow patch.
Video and screenshot review
Upload a short product demo or a set of UI screenshots and get a written summary of what happens and what looks broken.
Cross-source research briefs
Combine many sources into a structured brief with citations, then return it as JSON for downstream tooling.

Prompt examples

Here are three quarterly filings. Build a table of revenue by segment, flag any restated numbers, and list the notes that explain them.

Watch this three-minute screen recording of the checkout flow and list every step where the interface behaves unexpectedly.

Given this stack trace and the attached modules, find the root cause and write the smallest patch with a regression test.

This model has conditional pricing. Monthly token totals alone cannot produce a reliable estimate; use the detailed pricing for the request specification, cache, and applicable time window.

FAQ

What is Step 5 Preview?

It is StepFun's new flagship model for agentic work. StepFun states it delivers frontier-level performance on software engineering and professional knowledge work, with particular strength in finance. It reads text, images, and video, and it is labeled a preview.

What input types does Step 5 Preview accept?

Text, images in JPG, PNG, WebP, or GIF, and video in MP4, QuickTime, or Matroska. StepFun's documentation allows up to 60 images in one request and video files under 128 MB, with best results on videos shorter than five minutes.

Does Step 5 Preview support structured output and tools?

Yes. StepFun documents tool calling, streaming, JSON mode, JSON Schema, and prompt caching, and you can set reasoning effort to low, medium, or high. Use these together for agents that must return validated data.

Why is it called a preview?

StepFun released it ahead of a final version, so the model may be updated and results can shift between revisions. Treat it as a strong candidate for evaluation and keep a fallback model configured.

How does it compare with Step 3.7 Flash?

Step 5 Preview is the larger flagship with video input and a far longer context window, built for harder professional tasks. Step 3.7 Flash is the lighter model for high-volume coding and search loops where speed matters more.

How much does Step 5 Preview cost?

On TokenLab, Step 5 Preview costs Input $1.00 / Output $2.70 per 1M tokens. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.

What are the context window and output limit of Step 5 Preview?

Step 5 Preview accepts up to 1,000,000 tokens of context and returns up to 64,000 tokens in one response.

Which endpoint should Step 5 Preview use?

Use https://api.tokenlab.sh/v1/messages for Step 5 Preview. The request example below shows the matching code shape.

Which operations does Step 5 Preview support?

Step 5 Preview supports Text to text. Select an operation above to see its endpoint and request example.