Reportedly: China Telecom Open-Sources Xing4.0-29B-Intent to Turn Commands Into Structured Parameters
Reported by Pandaily · not yet confirmed by the company or a second independent outlet. We update this page when it is.
China Telecom's XingChen lab released Xing4.0-29B-Intent, which classifies intents and fills structured arguments in one JSON output. It includes 2.75 million training samples and an inference SDK, Pandaily reported.
China context
- Original name
- 星辰 Xing4.0-29B-Intent
- Outside China
- Open weights · pandaily.com
- Claims
- Company-reported; not yet independently evaluated
- For builders
- Builders outside China can use the open weights and SDK to implement intent parsing without building their own training pipeline.
- For investors
- Investors can monitor whether this model gains traction in developer communities as a signal of China Telecom's competitiveness in specialized AI models.
China Telecom's XingChen lab released Xing4.0-29B-Intent, a model that classifies intents and fills structured arguments in one JSON output. The release includes 2.75 million training samples and an inference SDK.
The model performs intent classification and parameter extraction in a single JSON output, suggesting a unified sequence-to-sequence approach. The 2.75 million training samples indicate a substantial dataset for intent-related tasks.
Developers outside China gain access to a specialized intent-parameter extraction model with open weights, potentially reducing the cost of building conversational AI systems that require structured command parsing.
The open-sourcing of a specialized intent model with training data and SDK can lower development costs for companies building voice assistants, chatbots, or automation tools that need to convert natural language commands into structured parameters.
Observable next signals include whether the model is adopted in open-source frameworks or integrated into commercial products, and whether China Telecom releases larger or domain-specific variants.