Hunyuan Large
Hunyuan Large is Tencent's open-source flagship language model at 389 billion parameters, one of the largest publicly released models in the industry. It uses a Mixture-of-Experts (MoE) architecture that activates only a subset of parameters per token, keeping inference costs manageable despite the enormous total parameter count. Tencent released Hunyuan Large to establish a credible open-source frontier model for Chinese enterprise and research use.
The model excels at knowledge-intensive tasks, long-document understanding, bilingual Chinese-English reasoning, and general instruction following. Its scale gives it broad factual coverage and strong few-shot performance across diverse domains.
Key Features
389B total parameters via Mixture-of-Experts architecture
Strong bilingual performance in Chinese and English
Broad factual coverage from large-scale pretraining
Efficient per-token compute via sparse MoE activation
Open-source weights available for research and enterprise self-hosting
Long-document comprehension and multi-turn dialogue
Ideal Use Cases
Enterprise knowledge management and document understanding in Chinese and English
Research applications requiring open-weight access to frontier-scale models
Long-form content generation and summarization
Complex question answering across broad knowledge domains
Example Prompts for Hunyuan Large
Technical Specifications
| Provider | Tencent |
| Category | Text |
| Modality | Text -> Text |
Frequently Asked Questions
Try Hunyuan Large now
Start using Hunyuan Large instantly — 100 free credits, no credit card required. Access 750+ AI models through one platform.