---
id: "stepfun/step-5-preview"
also: ["stepfun/step-5-preview-20261008", "stepfun-ai/Step-5-Preview", "requesty/stepfun/step-5-preview", "vercel/stepfun/step-5-preview"]
card_usd_per_m: "1.0 in, 2.7 out, at stepfun/fp8"
cheapest_list_usd_per_m: "1.0 in, 2.7 out, at featherless/fp4 on openrouter"
cheapest_effective_usd_per_m: "0.0976 in, 2.7 out, at deepinfra on requesty"
context: 1000000
max_output: 64000
input: ["text", "image", "video"]
output: ["text"]
released: "2026-10-08"
reasoning: ["high", "medium", "low"]
api: ["/api/v1/chat/completions", "/api/v1/completions", "/api/v1/messages", "/api/v1/responses"]
params: ["frequency_penalty", "include_reasoning", "logprobs", "max_tokens", "min_p", "presence_penalty", "reasoning", "reasoning_effort", "repetition_penalty", "response_format", "seed", "stop", "structured_outputs", "temperature", "tool_choice", "tools", "top_k", "top_logprobs", "top_p"]
endpoints: 6
gateways: "vercel (3), openrouter (2), requesty (1)"
no_training: 4
trains: 2
retains_prompts: 4
intelligence: 28.5000
---
# StepFun: Step 5 Preview

Step 5 Preview is StepFun’s flagship AI model for agentic coding, professional knowledge work, and financial analysis. Built on a sparse Mixture-of-Experts architecture with 600B total parameters and 27B active per token, it combines a 1M-token context window with vision input. From building interactive applications and debugging code to conducting research and producing analytical reports, it supports complex workflows that require sustained reasoning, tool use, and iterative refinement.