"At present, the capabilities of many agents are still at the elementary level. The direction of industry development is to truly realize agents with stronger reflection, planning and autonomous decision-making capabilities." On April 17, at the Volcano Engine 2025 FORCE LINK AI Innovation Tour in Hangzhou, Tan Dai, president of Volcano under ByteDance, said in an interview. Tan Dai believes that the deep thinking model is the basis for building an agent. The model must be able to think, plan and reflect well, and must support multimodality, just like humans have vision and hearing, so that the agent can better handle complex tasks. On the same day, Doubao 1.5 deep thinking model was officially released. According to reports, in addition to outstanding performance in general tasks such as mathematics, programming, scientific reasoning, creative writing, etc., this model also launched a visual version of the deep thinking model with visual reasoning ability, which can associate and think about what it sees like humans.