Hy4 Preview
Neural Network
Hy4 Preview is a Tencent model for coding agents and complex tool-calling scenarios. Access the Hy4 Preview API via BotHub.
Max answer length
(in tokens)
Context size
(in tokens)
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works Hy4 Preview?
Hy4 Preview is a Tencent model based on the mixture-of-experts architecture: 770 billion parameters total, with 49 billion active per request. It is designed for coding agents, complex tool-calling scenarios, and productivity tasks, making it ideal for multi-step chains requiring function calls, result analysis, and continued execution. Connect to Hy4 Preview via BotHub without a VPN or foreign card: pay in rubles with a Russian card only for tokens used, which never expire. Access over 250 models in one window for easy comparison, and use a unified OpenAI-compatible API to swap models without rewriting integrations. Companies get contracts, invoices, EDI, and an admin panel with limits; chats are AES-GCM encrypted and not saved. Useful for building agents that edit repositories and run tests, debugging legacy code, connecting the model to internal services via function calling, and breaking down technical specifications into actionable steps.Frequently asked questions about Hy4 Preview
You can use generated results for commercial purposes. You own all rights to the content you create. The only restriction: make sure your prompt does not include copyrighted third-party material. You are responsible for respecting the rights to any input data.
hy4-preview is a Tencent text model with reasoning before answering. It is suitable for working with long documents, analyzing and structuring information, writing and parsing code, and STEM-related tasks. A context window of up to 1,048,576 tokens allows you to upload large materials in their entirety and analyze them in a single dialogue.
The model generates up to 64,000 tokens per response, which is dozens of pages. This is enough for a large technical document, a detailed analysis, or an entire large code module. If the material is even larger, break the task into parts: the huge context window allows you to keep the entire project in one dialogue.
Yes, the reasoning mode is confirmed: the model first builds a chain of steps and only then formulates an answer. This helps with math, logic, code debugging, and multi-step document analysis. The difference is small for simple queries, but for complex ones, the chance of getting a correct result immediately increases.
Yes, both. hy4-preview can call functions and provide structured output in JSON, making it easy to integrate into agents and workflows. In BotHub, it is available via a unified OpenAI-compatible API, so you can switch to it without rewriting your integration.
In addition to text, the model accepts documents: upload a file and ask for a summary, version comparison, or data extraction into a table. We have no data on image, audio, or video input. For such formats, BotHub offers multimodal models in the same window and via the same API.
We do not have specific data on the quality of Russian for this model, so we cannot make any promises—it is easier to test it with your own task. It is inexpensive: BotHub charges per token used, with no subscription, no VPN, and no foreign card required, and you can compare the response with another model right next to it.