BEIJING, Jan 22 – Chinese tech giants ByteDance and DeepSeek are ramping up the global AI race with the launch of powerful new reasoning models aimed at challenging Microsoft-backed OpenAI’s dominance in advanced artificial intelligence.
ByteDance, the owner of TikTok, has unveiled Doubao-1.5-pro, an upgraded AI model that the company claims outperforms OpenAI’s “o1” in AIME, a benchmark for evaluating how well AI systems process and execute complex instructions. The release reflects ByteDance’s growing ambition to stake its claim in the highly competitive AI space.
DeepSeek, a rising star in China’s AI sector, has introduced DeepSeek-R1, an open-source reasoning model it says rivals OpenAI’s performance on multiple benchmarks. This follows recent tests that showed DeepSeek’s V3 large language model surpassing those of OpenAI and Meta, despite operating on a smaller development budget. The company has gained global attention for offering advanced AI capabilities at significantly lower prices.
Both companies are pushing aggressively on pricing to disrupt the market. DeepSeek offers access to its models at 16 yuan ($2.20) per million tokens, a dramatic drop from OpenAI’s 438 yuan fee for similar usage. ByteDance has gone even further, pricing Doubao-1.5-pro-32k at 2 yuan per million tokens and the more powerful Doubao-1.5-pro-256k at 9 yuan via its Volcano Engine cloud platform.
These developments come as OpenAI prepares to launch its latest reasoning model, o3 mini, in the coming weeks. The global race for AI innovation, triggered by OpenAI’s launch of ChatGPT in late 2022, has intensified, with Chinese firms like ByteDance, DeepSeek, Moonshot AI, and iFlyTek entering the fray. By combining cutting-edge performance with cost-effective solutions, these companies are positioning themselves as formidable competitors to OpenAI and other established players in the rapidly evolving AI industry.
