Tencent has announced the open-source release of four lightweight large language models (LLMs) under its Hunyuan ecosystem. It includes 0.5B, 1.8B, 4B, and 7B parameters. These models are said to bring advanced AI capabilities to low-power environments. The company shared the update via its official X handle, highlighting the models’ broad compatibility and performance across several benchmark tests.
The new models are designed to run on a single consumer-grade GPU. This makes them well-suited for integration into smart vehicles, home devices, mobile phones, and personal computers. It also allows developers to access these models directly from GitHub and Hugging Face.
🚀We're expanding the Tencent Hunyuan open-source LLM ecosystem with four compact models (0.5B, 1.8B, 4B, 7B)! Designed for low-power scenarios like consumer-grade GPUs, smart vehicles, smart home devices, mobile phones, and PCs, these models support cost-effective fine-tuning… pic.twitter.com/CknskVqPem
— Hunyuan (@TencentHunyuan) August 4, 2025
One of the main highlights is the ability to switch between “fast thinking” for concise outputs and “slow thinking” for more in-depth, reasoning-heavy tasks. Besides, all four models support a native 256K token context window.
Moreover, the company has also shared some benchmark graphs showing the Hunyuan models outperforming or staying competitive with many models in their class. The Hunyuan 0.5B to 4B models show consistent gains across MMLU, BBH, GSM8K, and other reasoning benchmarks, with the 4B variant scoring particularly well in math (GSM8K – 87.5) and language tasks.
The Hunyuan 7B model holds its ground against OpenAI’s o1-mini and Qwen’s 7B/8B variants. It shows strong results in evaluations like MATH500 (93.7), DROP (85.9), and AIME 2024 (81.1), signaling its effectiveness across STEM, comprehension, and logic-heavy benchmarks.
Apart from that, all four models are compatible with popular deployment frameworks like SGLang, vLLM, and TensorRT-LLM. The models also support capabilities like tool calling, task planning, and decision-making.
That’s all about Tencent for now. But if you want to stay updated with all the latest tech and AI news, then do join us on WhatsApp, where we share the latest news, reviews, and more.
