Vanchin
StreamLake's enterprise model-serving and AI computing platform, offering hosted inference and model customisation over a curated hub of open and in-house models.
Per-million-token pricing in USD with split input/output rates listed on the product page itself; example given is DeepSeek-V4-Flash at 0.14 USD/M input and 0.28 USD/M output. No free tier is stated. Served models listed include DeepSeek-V4 Flash/Pro, GLM-5.2, Kimi-K2.7-Code, MiniMax-M3, Xiaomi MiMo-V2.5-Pro, Qwen and Kuaishou's own Keye-VL.
What it does
- Model inference
- Model hosting
- Model training
How it compares
-
Kimi
xcellent," and will be open-sourced globally. Reasoning 1000k Try Now Kimi-K2.7-Code Input ： $ 0.95 / M Tokens · Output ： $ 4 / M Tokens Kimi-K2
Official documentation · 5 Sep 2026
Sources
- Pricing
- Sold within
-
(pagetext, 2026-09-30). "Flexible Pricing with Transparent and Predictable Costs" table prices this product's inference, e.g. "DeepSeek-V4-Flash 1M tokens $0.1400 $0.2800 $0.0280"; page offers "Get Started" and "Console". Own rate card billed on the StreamLake account (not a parent offering). No free tier stated.
Also from Kuaishou
-
CodeFlicker Developer tool
Kuaishou/StreamLake's AI code editor, providing autonomous coding agents, context-aware completion, Figma-to-code conversion and IDE extensions.
-
CyberCut Application
StreamLake's AI video studio that cuts long videos into short-form clips, generates videos from scripts with AI voiceover and digital avatars, and supports text-based and conversational editing.
-
Kling AI Application
Kuaishou's web and mobile AI creative studio that turns text, images and references into generated video, images, sound and edits.
-
Kling AI Developer API Model API
Kling's open-platform REST API that lets developers call its video and image generation models programmatically from Singapore or Beijing endpoints.
Open Kuaishou in the directory
Something wrong here? Send a correction — quote this product id: kuaishou-vanchin.