Hunyuan-A13B Technical Report
Tencent releases Hunyuan-A13B, an open 80B MoE that activates 13B parameters and adds dual-mode chain-of-thought.
Hunyuan-A13B is an open-source Mixture-of-Experts language model with 80 billion total parameters and 13 billion active during inference. It was pretrained on a filtered 20-trillion-token corpus with enhanced STEM curation, then improved with supervised fine-tuning and large-scale reinforcement learning. A dual-mode chain-of-thought uses fast thinking for routine queries and slow thinking for multi-step problems. Reported results are competitive with much larger models on mathematics, science, programming, language understanding, and agent tasks.
- Hunyuan-A13B is an open MoE with 80B total and 13B active parameters.
- Pretraining used a filtered 20-trillion-token corpus with extra STEM curation.
- Dual-mode chain-of-thought switches between fast and slow reasoning.
- It is competitive on math, science, programming, language, and agent tasks.
- Weights are released for open research and practical deployment.
Full article135 words · extracted from huggingface.co · click to collapse
We present Hunyuan-A13B, an open-source large language model based on a Mixture-of-Experts architecture. It contains 80 billion total parameters but activates only 13 billion during inference, balancing model capability, computational efficiency, and deployment cost. The model is pretrained on a rigorously filtered 20T-token corpus with enhanced STEM data curation, improving factual reliability and reasoning ability. High-quality supervised fine-tuning and large-scale reinforcement learning further enhance its overall performance. Hunyuan-A13B also introduces a dual-mode Chain-of-Thought framework that adapts reasoning depth to task complexity: fast thinking for routine queries and slow thinking for complex, multi-step problems. Evaluations show competitive performance across mathematics, science, programming, general language understanding, and agent tasks, often approaching that of much larger models. Its high inference throughput makes it suitable for latency-sensitive applications. We release Hunyuan-A13B to support open research and practical LLM deployment.
Text extracted automatically; images, tables and formatting may be missing. Original: https://huggingface.co/papers/2609.27284