StepFun Yuewen 阶跃星辰 跃问
StepFun's multimodal chat app, backed by a $717M round and IPO plans
Overview
Yuewen (跃问) is StepFun's (阶跃星辰) free consumer AI assistant — multimodal chat, document understanding, image/video generation and voice — powered by the Step model family, alongside a developer platform (Step Plan) selling agentic-coding API access from $6.99/month.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 12 Jul 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- 100 prompts / 5 hours (~1,500 requests)
- Flagship model + smart routing
- Multi-device login
- 400 prompts / 5 hours (~6,000 requests)
- Priority API access and support
- 1,500 prompts / 5 hours (~22,500 requests)
- 5,000 prompts / 5 hours (~75,000 requests)
- +40% Studio Creation Credits
What you can produce with StepFun Yuewen 阶跃星辰 跃问
- Multimodal chat: text, document understanding, image/video generation, voice
- Step 3.5 Flash + DeepSeek-R1 reasoning fallback
- Step Plan developer API subscription for coding agents (Cursor, Windsurf, Cline)
- Step-Video-T2V (204-frame, 30B-parameter text-to-video)
- Step-Audio 2 speech understanding/conversation model
- Step1X-Edit open-source image editing model
- 25+ open-source repos under Apache 2.0 (stepfun-ai org)
ASEAN Perspective
StepFun Yuewen 阶跃星辰 跃问 in Southeast Asia
ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).
StepFun built its name on multimodal strength rather than pure chat — Step3-VL-10B reportedly matches open models 10-20x its size, and Step-Video-T2V, Step-Audio 2 and Step1X-Edit are all credible, Apache-2.0-licensed releases with real GitHub activity (25 repos). Yuewen, the consumer wrapper around Step 3.5 Flash, is free and folds in DeepSeek-R1 for reasoning-heavy queries — a sensible way to cover a capability gap without training it in-house.
The January 2026 Series B+ (over 5 billion yuan, roughly $717M — reportedly the largest single China large-model round in months) and February reports of a Hong Kong IPO exploration signal real momentum, following MiniMax and Zhipu's earlier HK listings. The Step Plan developer subscription ($6.99 to $99/month across four tiers) targets Cursor/Windsurf-style coding workflows directly, competing on price against Zhipu's GLM Coding Plan. Brand recognition outside China remains low relative to Doubao or Kimi.
What people say
StepFun (阶跃星辰) is a Shanghai lab founded in April 2023 by Jiang Daxin, previously a Global VP and chief scientist at Microsoft's Software Technology Center Asia — a research pedigree that shows up most clearly in the company's multimodal releases rather than its consumer chat app. Step3-VL-10B, a 10-billion-parameter multimodal model, is pitched by the company as matching open-source models ten to twenty times its size; Step-Audio 2 handles industry-strength speech understanding and conversation; Step1X-Edit is an open-source image editor benchmarked against GPT-4o and Gemini 2 Flash; Step-Video-T2V generates up to 204 frames from a 30B-parameter text-to-video model. Most of this ships on GitHub under Apache 2.0 across 25 repositories in the stepfun-ai org — an unusually open posture for a Chinese frontier lab.
Yuewen (跃问) is the free consumer face of that research: a multimodal assistant on iOS, Android and web doing text chat, document understanding, image/video generation and voice interaction, currently running on Step 3.5 Flash with DeepSeek-R1 folded in for reasoning tasks the house model handles less well. It's a sensible way to plug a gap without a from-scratch reasoning model, though it does mean part of the product depends on a rival lab's weights.
The money story moved fast in early 2026. StepFun closed a Series B+ of more than 5 billion yuan (about $717 million) in January — reportedly the largest single round in China's large-model sector in the preceding months — and Bloomberg reported in February that the company was exploring a Hong Kong IPO that could raise roughly $500 million, following MiniMax and Zhipu AI onto the exchange as the third of the 'Six Little Tigers' to head that direction.
On the developer side, StepFun launched Step Plan in 2026: four subscription tiers (Flash Mini, Flash Plus, Flash Pro, Flash Max) running from $6.99 to $99 a month, billed by 'prompts' rather than raw requests (roughly 15-20 requests per prompt) on a rolling 5-hour window — 100 prompts on Flash Mini scaling to 5,000 on Flash Max. It's positioned squarely at Cursor, Windsurf and Cline users and undercuts most Western coding-assistant subscriptions on price, though the two middle tiers' exact pricing wasn't published at the time of writing.
Summary of public user & expert reviews, compiled by RECATOOLS.
About this listing
This entry was compiled from publicly available data including StepFun Yuewen 阶跃星辰 跃问's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with StepFun Yuewen 阶跃星辰 跃问 unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to StepFun Yuewen 阶跃星辰 跃问 directly →
Spotted something out of date? Suggest an update →
Alternatives to StepFun Yuewen 阶跃星辰 跃问
More in LLMs & Chat
