Xing4.0 29B A4B
Neural Network
Xing4.0 29B A4B is an open TeleAI language model with a mixture-of-experts architecture for text tasks.
Specifications
- Input
- text, documents
- Output
- text
- Context size
- 262 144 tokens
- Max answer length
- 235 929 tokens
- Reasoning effort
- low, medium, high
- Capabilities
- reasoning, function calling
How it works Xing4.0 29B A4B?
Xing4.0 29B A4B is an open-weights language model by the TeleAI team, the AI division of China Telecom, and a continuation of the TeleChat line. It is based on a mixture-of-experts architecture: out of 29 billion parameters, about 4 billion are active per token, making the model more efficient than dense networks of comparable size while retaining the knowledge volume of a large network. Via BotHub, it is accessible without a VPN or foreign card: you pay with a Russian card in rubles only for the tokens used, which do not expire. With over 250 models in one window, it is easy to compare answers with other neural networks, and data is transmitted with AES-GCM encryption and not stored. Developers have access to a unified OpenAI-compatible API: you can switch to another model by changing the identifier without rewriting the integration, and companies get contracts, invoices, electronic document management, and an admin panel with limits. Use Xing4.0 29B A4B for drafting emails and descriptions, summarizing long documents and correspondence, preparing standard support responses, and as an economical layer in a multi-model system where simple requests go to a lightweight model and complex ones to a heavy one.Frequently asked questions about Xing4.0 29B A4B
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.
This is a text model with reasoning and tool support. It is suitable for long texts, document analysis and parsing, coding and debugging, analytics, and multi-step tasks. A context of up to 262,144 tokens allows you to keep voluminous materials in the dialogue in their entirety.
According to our data, the output limit is 235,929 tokens per response. This is enough for a voluminous report, detailed documentation, a large block of code, or a detailed analysis that does not need to be assembled from many short fragments.
Yes, the reasoning mode is confirmed by our data. The model first builds a solution path and then provides an answer — this helps in mathematics, logic, STEM tasks, and code analysis, where the sequence of steps is important, not a quick guess.
Yes, the model supports function calling and structured JSON output. You can connect it to your tools, databases, and external services via the unified OpenAI-compatible BotHub API and receive predictable responses that are immediately parsed by code.
According to our data, the model accepts text and documents as input — they can be sent for analysis, summarization, and comparison. If you need to work with images, audio, or video, multimodal models are available nearby in BotHub: switching takes a couple of seconds.
We do not have data on the quality of work with specific languages, so it is more reliable to check with your own tasks. In BotHub, it is simple: pay-as-you-go for tokens, no subscription, no VPN, and no foreign card required — compare the result with other models in one window.