The tech giant’s latest offering leverages large-scale reinforcement learning, rivalling DeepSeek in top benchmark tests.
Tencent launched its Hunyuan T1 AI reasoning model, which rivals DeepSeek's R1 in performance and price, leveraging large-scale reinforcement learning and a hybrid Transformer-Mamba architecture. T1 outperformed DeepSeek-R1 on the MMLU Pro benchmark but trailed OpenAI's o1; it excelled in Chinese language capabilities. T1's pricing is similar to DeepSeek's, costing 1 yuan per million input tokens and 4 yuan per million output tokens.
