Skip to content
ShilpBot
AI Models

StepFun Step 5 Preview: Pricing, Specs and Limits

StepFun Step 5 Preview: sparse mixture-of-experts flagship for agentic work, 1M context, 1.00 in / 2.70 out per 1M tokens. Specs, limits and test prompts.

By ShilpBot Team · Published · 3 min read

Step 5 Preview is StepFun's flagship model for agentic work. It uses a sparse mixture-of-experts architecture with 27 billion active parameters out of 600 billion in total, and the listing says it performs strongly in software engineering and professional tasks.

Its output price of $2.70 per million tokens is low for a flagship, while its output cap of 64,000 tokens is lower than many rivals.

Prices and specifications checked against the public OpenRouter catalogue on 2026-11-16. See the live model page for current limits.

StepFun Step 5 Preview quick facts

DetailValue
DeveloperStepFun
Model IDstepfun/step-5-preview
Added to the catalogue2026-10-08
TypeSparse mixture-of-experts flagship for agentic work
Parameters27B active / 600B total
Context windowabout 1M tokens (1,000,000)
Maximum output64,000 tokens
Input typestext, image, video
Price1.00 in / 2.70 out per 1M tokens (cached input 0.10)

Feature support from the catalogue: Tool calling: listed; JSON / structured output: listed; Adjustable reasoning effort: listed.

What StepFun Step 5 Preview is good for

  • Agentic software engineering at a lower output price than many flagships.
  • Image and video understanding alongside text.
  • Professional tasks such as report drafting.

Limits of StepFun Step 5 Preview

  • "Preview" means the model can change before the final release.
  • The 64,000-token output cap limits very long single responses.
  • Run your own checks before depending on benchmark claims.

How StepFun Step 5 Preview compares

  • Tencent Hy4: It has a smaller context window (1,000,000 against 1,048,576 tokens); its output cap is 64,000 tokens against 65,536; it also accepts video input; it costs 1.00 input and 2.70 output per million tokens against 0.42 and 1.25.
  • OpenAI GPT-6.1 Terra: It has a smaller context window (1,000,000 against 1,050,000 tokens); its output cap is 64,000 tokens against 128,000; it also accepts video input; it does not accept file input; it costs 1.00 input and 2.70 output per million tokens against 1.00 and 5.00.
  • Nex AGI Nex-N2.6: It has a larger context window (1,000,000 against 262,144 tokens); its output cap is 64,000 tokens against 235,929; it also accepts video input; it costs 1.00 input and 2.70 output per million tokens against 0.12 and 0.40.

Calling StepFun Step 5 Preview from code

Use the model ID stepfun/step-5-preview with any OpenAI-compatible client pointed at OpenRouter:

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "stepfun/step-5-preview", "messages": [{"role": "user", "content": "Explain what you can do in two sentences."}]}'

Prompts that show what StepFun Step 5 Preview can do

  1. Give it a feature request and ask for an implementation plan before code.
  2. Upload a design mock-up and ask for component structure.
  3. Compare cost for a 50-step agent run against GPT-6.1 Terra.

What StepFun Step 5 Preview costs

The catalogue lists 1.00 per million input tokens and 2.70 per million output tokens, with cached input at 0.10 per million. Sending 1 million input tokens and receiving 100,000 output tokens would cost about 1.27. A single coding-agent step with a 150,000-token prompt and a 10,000-token answer would cost about $0.18 before caching. If your client reuses the same prompt prefix, cached input at 0.10 per million tokens can cut the input side of that bill by about 90 percent. Prices are set by the provider, can change, and may differ between hosting providers, so check the live model page before you budget.

Keep reading

Compare every paid model by price in our AI model pricing roundup, or see the free models roundup. New to terms like mixture-of-experts? Start with how neural networks learn.

Frequently asked questions

Is Step 5 a final release?

The listing calls it a preview, so expect changes before a final version.

What are the cost and limits of StepFun Step 5 Preview?

As of 2026-11-16 the catalogue lists 1.00 per million input tokens and 2.70 per million output tokens. About 1M tokens (1,000,000) context window, up to 64,000 output tokens, model ID stepfun/step-5-preview. Prices can change, so check the live model page.

#StepFun Step 5 Preview#StepFun model#AI model pricing#AI API pricing#AI chat model#agentic model#OpenRouter