China’s DeepSeek challenges Google DeepMind and OpenAI with new AI model
DeepSeek’s V3.2-Speciale equals Google’s Gemini 3 Pro in reasoning capabilities, the Chinese start-up says
The achievement by the open-source lab has sparked extensive discussion within the AI research community as it coincides with the prestigious annual Conference on Neural Information Processing Systems, or NeurIPS.
In its announcement on Monday, Hangzhou-based DeepSeek said V3.2-Speciale equalled Google’s Gemini 3 Pro, released two weeks ago, in reasoning capabilities, while V3.2-Speciale’s base model, the V3.2 introduced on the same day, performed on par with OpenAI’s GPT-5 launched in August.
DeepSeek said that V3.2-Speciale achieved gold-medal performance on the International Mathematical Olympiad test – an accomplishment previously reached only by internal models from OpenAI and Google DeepMind that have not been made public.
On social media, Google DeepMind principal research engineer Susan Zhang praised DeepSeek for the detailed technical report accompanying its new models, noting the start-up’s efforts in stabilising the models post-training and enhancing their agentic capabilities.
Shares of Google’s parent company, Alphabet, fell 1.65 per cent on Monday.