Your cookie choices

We use necessary cookies to make this site work. With your permission, analytics cookies help us understand usage and marketing cookies support advertising.

You can reject non-essential cookies and still use the site. Change your choices anytime in Cookie settings.

Cookie settingsCookie policy

StepFun (Jieyue Xingchen) review

Updated September 28, 2026 · DeployCue Editorial Team · Facts checked September 26, 2026

We earn a commission when you sign up through links on this site. Partner status can affect where a brand appears in our rankings. Partners with a live offer in your country come first, then every other brand by score.

Our verdict

StepFun (Jieyue Xingchen) suits developers who want text, speech and image models behind one OpenAI-compatible endpoint and who can build on a provider based in China. Entry input fees start at CNY 0.70 to CNY 7 and output fees at CNY 2.10 to CNY 20, while step-5-preview takes a 1,000,000 context window.

Key facts

Best for
Teams wanting text, speech and image models on an OpenAI-compatible API from a China-based provider.
Input price from
CN¥0.70 to CN¥7↗
Output price from
CN¥2.10 to CN¥20↗
Models offered
Step-5-preview, Step-3.7-flash, Step-3.5-flash, Step-3.5-flash-2603, Step-1o-turbo-vision, Stepaudio-3-realtime-preview, Stepaudio-3-chat-preview, Stepaudio-2.5-realtime, Stepaudio-2.5-chat, Step-1o-audio, Step-audio-2, Step-audio-r1.5, Stepaudio-3-tts, Stepaudio-2.5-tts, Step-tts-2, Step-tts-mini, Stepaudio-3-asr-max, Stepaudio-2.5-asr, Step-2x-large, Step-image-edit-2↗
Largest context window
1,000,000↗
OpenAI-compatible API
Yes↗
Legal entity
Shanghai StepFun Intelligent Technology Co., Ltd.↗
Headquarters
China↗
Support channels
Email, Community↗
Official website
stepfun.com

Facts checked September 26, 2026, 3 sources

Apps and profiles

Elsewhere
WikidataWikipediaXgithub.comhuggingface.co

Pros and cons

Pros

  • One catalogue covers text, speech, audio and image work via step 5 preview, step 3.7 flash and step 3.5 flash
  • step-5-preview reaches a 1,000,000 context window
  • An OpenAI-compatible API, so existing clients and SDKs carry over
  • Entry input fees run CNY 0.70 to CNY 7 and output fees CNY 2.10 to CNY 20

Cons

  • Support runs on email and community only, so urgent questions wait on a reply
  • Prices are quoted in CNY from a provider headquartered in China, which adds currency and jurisdiction friction
  • Entry input fees rank 50 of 71 on our data, with entry output fees at 42 of 70
StepFun (Jieyue Xingchen)

Ready to try StepFun (Jieyue Xingchen)?

Scored 8.9 / 10 by DeployCue.

See alternatives

Who it suits

  • Developers already sending OpenAI-style requests who want speech and image models in the same catalogue
  • China-based teams billing in CNY who want text, speech and image work behind one endpoint
  • Small teams prototyping long-context work on step-5-preview

Who should look elsewhere

  • Teams that need live chat or phone support, since support runs on email and community
  • Buyers who price only in USD and would rather route through a gateway such as Salad at $0.06
  • Teams that want a wide general-purpose catalogue rather than one house family, as Azure AI Foundry offers

Core features

The catalogue in step 5 preview, step 3.7 flash and step 3.5 flash spans step-5-preview and step-3.7-flash for text, step-1o-turbo-vision for vision, the stepaudio and step-tts families for speech and transcription, and step-image-edit-2 for image work. step-5-preview carries the largest window on the platform at 1,000,000. The endpoint is an OpenAI-compatible API, so most request code carries over.

How it compares on our data

Context is the stronger side: the largest window on the platform puts StepFun at 21 of 72 brands we track on that measure. Price is the weaker side. Entry input fees sit at 50 of 71, and entry output fees at 42 of 70.

How StepFun (Jieyue Xingchen) compares

Frequently asked questions

What does it cost to call StepFun models?+

Entry input fees start at CNY 0.70 to CNY 7 and entry output fees at CNY 2.10 to CNY 20, both quoted in CNY. Those are the cheapest listed rates, and the range widens across the catalogue, so the model you pick moves the bill. Check the per-model table before you commit a large workload.

How large a context window can I use?+

step-5-preview takes a context window of 1,000,000, the figure behind a rank of 21 among 72 brands we track on that measure. Long documents and large codebases can go in one request rather than being split into chunks.

Is the API OpenAI-compatible?+

Yes. The platform offers an OpenAI-compatible API, so existing clients and SDKs can point at the endpoint. Model names differ from OpenAI's, and the catalogue in step 5 preview, step 3.7 flash and step 3.5 flash is a house range, so check the model list before migrating a client.

Which models can I call?+

The catalogue in step 5 preview, step 3.7 flash and step 3.5 flash covers text with step-5-preview and step-3.7-flash, vision with step-1o-turbo-vision, speech and transcription through the stepaudio and step-tts families, audio understanding, and image work with step-image-edit-2. Everything sits behind the same OpenAI-compatible API.

Where is the provider based?+

The operator is Shanghai StepFun Intelligent Technology Co., Ltd., and the headquarters country is CN. Prices are quoted in CNY, so teams billing in other currencies should weigh conversion and the practical side of serving traffic from a China-based provider.

How do I get support?+

Support runs on email and community, which is email and a community forum. That suits developers comfortable reading documentation and asking peers, and it means a time-sensitive production question may wait for a reply rather than a live response.

How do its entry prices compare with other providers?+

Entry input fees rank at 50 of 71 brands and entry output fees at 42 of 70, while the context window of 1,000,000 ranks at 21 of 72. The trade-off is window size against entry rate.

Sources

Compiled from 3 source pages, checked September 26, 2026. We did not open an account.

StepFun (Jieyue Xingchen)

StepFun (Jieyue Xingchen)

Teams wanting text, speech and image models on an OpenAI-compatible API from a China-based provider.

8.9/10See alternatives