stepfun/step-5-preview

模型发布 OpenRouter 模型 原文语言:中文

OpenRouter 于 2026-10-08 收录该模型(收录时间不等于模型正式发布日期);上下文:1,000,000 tokens;输入价:$1.0000/百万 tokens;输出价:$2.7000/百万 tokens;定位(来源描述):Step 5 Preview is StepFun's flagship model for agentic work, built on a sparse Mixture-of-Experts…

全文

StepFun: Step 5 Preview

stepfun/step-5-preview

Model weights

Compare

Step 5 Preview is StepFun's flagship model for agentic work, built on a sparse Mixture-of-Experts architecture (27B active / 600B total parameters). It performs strongly in software engineering and professional knowledge work, with particular strength in finance.

It is designed for extended tasks that span large codebases and documents, using tools and refining results over multiple steps.

Modalities

In / out price

$1 / $2.70per 1M

Context

1.0M

Released

Oct 8, 2026

Compare

Providers

Different companies host the same model. OpenRouter routes your request to one of them based on the routing mode you pick: Balanced (price + speed), Nitro (fastest), or Floor (cheapest).

Latency / throughput

$1.00$2.70$0.101.76s138 tok/s

99.87%

$1.00$2.70$0.1013.00s8 tok/s

84.22%

ProviderInput / Output /M

Input $1.00 / Output $2.70

138 tok/s99.87% uptime

Input $1.00 / Output $2.70

8 tok/s84.22% uptime

Pricing

The average price customers actually pay for this model, next to the prices providers post. Caching and discounts mean the price actually paid is often well below the listed one.

Weighted average

Input

$0.1449/M tokens

Output

$2.70/M tokens

Step 5 Preview — Price history

Price history

No effective price data for this selection

$0.145$2.7095.0%100.0%

$0.148$2.7094.7%0.0%

ProviderInput / Output /M

Token share

Effective in $0.145 / Effective out $2.70

Token share

100.0%

Effective in $0.148 / Effective out $2.70

Token share

0.0%

Performance

Throughput is how fast the model writes (tokens per second — higher is better). Latency is total round-trip time (lower is better). TTFT is time-to-first-token — how long before you see anything appear (lower is better).

Throughput

138tok/s

P50, best across providers

Latency

1.76s

P50, best provider

Reasoning effortLatency / throughput

Uptime

Uptime is the percentage of the past 3 days that at least one provider was responding to requests. Availability is the percentage of time that inference was successfully served. OpenRouter continuously monitors and uses the next-best provider when one returns an error.

Uptime (3d)The model was reachable. Request routed to a provider.

100.00%

Availability (3d)The model returned inference from any provider. Errors and empty responses count against it.

98.88%

Availability over the last 3 days

Last 72 hours

Availability 98.88%

3 Days Ago2 Days AgoYesterdayNow

Availability over the last 24 hours

OpenRouter Availability

98.83%

Without Routing

97.06%

When an error occurs in an upstream provider, we can recover by routing to another healthy provider, if your request filters allow it. You can access per-provider uptime data programmatically through the Endpoints API. Learn more about our load balancing and provider routing options.

Benchmarks

Scores on standardized evaluations and this model’s own outputs on shared benchmark prompts. Higher percentages are better — and each rank shows where this model lands among the models evaluated on that source.

Benchmark score summary for StepFun: Step 5 Preview (Artificial Analysis and Vals AI)

SourceBenchmarkScore

Artificial AnalysisStep 5 Preview Intelligence Index28.5

Artificial AnalysisStep 5 Preview HLE46.5%

Artificial AnalysisStep 5 Preview AA-LCR88.3%

Artificial AnalysisStep 5 Preview GDPval-AA54.5%

Artificial AnalysisStep 5 Preview CritPt20.9%

Artificial AnalysisStep 5 Preview SciCode58.9%

Artificial AnalysisStep 5 Preview Terminal-Bench 4.033.3%

Artificial AnalysisStep 5 Preview AA-Omniscience Accuracy36.4%

Artificial AnalysisStep 5 Preview AA-Omniscience Non-Hallucination Rate68.7%

Vals AIStep 5 Preview BioMysteryBench67.4%

Vals AIStep 5 Preview Code Migration v1.147.0%

Vals AIStep 5 Preview EMB v1.169.6%

Vals AIStep 5 Preview Finance Agent v2.158.9%

Vals AIStep 5 Preview Vals Finance Index55.4%

Vals AIStep 5 Preview HLAB-Vals-1.112.9%

Vals AIStep 5 Preview IOI38.1%

Vals AIStep 5 Preview Vals Legal Index70.2%

Vals AIStep 5 Preview Legal Research Bench40.4%

Vals AIStep 5 Preview MysteryMechanism18.5%

Vals AIStep 5 Preview ProofBench v1.142.0%

Vals AIStep 5 Preview Public Benefits Bench v1.158.1%

Vals AIStep 5 Preview Vals Science Index29.6%

Vals AIStep 5 Preview SRE Bench0.4%

Vals AIStep 5 Preview SWE-bench Multimodal40.8%

Vals AIStep 5 Preview Tax Agent Bench27.3%

Vals AIStep 5 Preview Terminal-Bench 4.031.8%

Vals AIStep 5 Preview Terminal-Bench-Science2.9%

Vals AIStep 5 Preview Vals Index54.8%

Vals AIStep 5 Preview Vibe Code Bench v1.170.4%

Apps

Public apps that send the most traffic to this model. Good signal for what real production workloads look like — and a hint at which use cases this model is best suited for.

1.

OpenClaw

OpenClaw is an open-source AI agent that connects to your messaging apps and takes real actions on your behalf, from running commands and browsing the web to managing files and sending emails.

6.29Ttokens

2.

Hermes Agent

Hermes Agent is an open-source, self-improving AI agent by Nous Research that runs persistently with memory across sessions, and builds reusable skills from experience. It comes with 40+ built-in tools, including web search, browser automation, and vision, plus scheduled automations and subagents.

1.83Ttokens

3.

pi

There are many coding agents, but this one is yours.

2.88Btokens

4.

Kilo Code

Kilo Code is an open-source AI coding agent that works across VS Code, JetBrains, and CLI to help developers ship code faster with agentic workflows.

2.28Btokens

5.

Nous Research API

Nous Research is an AI research team focused on advancing open-source AI.

1.61Btokens

Rest of OpenRouter

670B

Top 5 Apps

131B

Activity

Token volume and request traffic to this model over time.

API

Drop-in code to call this model. OpenRouter's API is OpenAI-compatible — most SDKs work by just swapping the base URL. The only thing that changes between models is the model slug below.

Not currently available through EU in-region routing. In-region routing docs

Frequently asked questions

Step 5 Preview is StepFun's flagship model for agentic work, built on a sparse Mixture-of-Experts architecture (27B active / 600B total parameters). It performs strongly in software engineering and professional knowledge work, with particular strength in finance. It is designed for extended tasks that span large codebases and documents, using tools and refining results over multiple steps.

Step 5 Preview costs $1.00/M input tokens and $2.70/M output tokens, with separate rates for Cache Read at $0.10/M tokens.

Step 5 Preview has a 1,000,000 token context window. It supports up to 32,768 completion tokens.

Yes. Step 5 Preview accepts tools for function calling. It also supports structured outputs via a JSON schema in response_format.

Step 5 Preview accepts text, images, and video as input and returns text.

Step 5 Preview is served by 2 providers on OpenRouter: Featherless and StepFun. Requests are routed to the best available provider, with automatic failover to the others, and you can pin or exclude providers with provider routing.

Step 5 Preview was released on October 8, 2026.

正文以来源语言显示。

参考来源

← 返回每日简报