Chinese AI company DeepSeek has denied renewed market speculation that its next-generation AI model, DeepSeek-R2, will launch between August 15 and August 30, according to a person familiar with the company’s plans. Rumors about an R2 debut have circulated for months. In early 2025, reports suggested a March 17 release, which the company also refuted. DeepSeek has yet to announce a launch date or reveal technical details for the model. Reuters previously cited three sources as saying that DeepSeek was accelerating R2’s rollout after initially targeting May, aiming for improved code generation and reasoning in languages beyond English. In June, The Information said the company’s CEO Liang Wenfeng remained dissatisfied with the model’s capabilities, pushing for further development to boost performance. DeepSeek has not commented publicly on the latest rumors. [iThome, in Chinese]
Related
