Chinese tech giant ByteDance has rolled out a new “deep research” feature for its AI assistant Doubao, making it available for free on the app, web, and PC platforms. The new tool aims to help users tackle complex tasks by quickly integrating extensive information and generating detailed research reports or visualized web results.

The new tool aims to help users tackle complex tasks by quickly integrating extensive information and generating detailed research reports or visualized web results.

Additionally, the Doubao app offers a one-click function to convert report content into a podcast format for convenient listening.

🔗 Source: AI Base

🧠 Food for thought

1️⃣ ByteDance’s aggressive AI pricing strategy reshapes market dynamics

ByteDance’s launch of Doubao’s new research feature aligns with the company’s pattern of offering significantly lower pricing in the AI space.

The company’s Doubao Pro model costs as little as 0.0008 yuan (approximately 0.011 US cents) per 1,000-token prompt, which is 99.8% cheaper than OpenAI’s GPT-4 model at 0.42 yuan per 1,000 tokens 1.

This pricing strategy is part of ByteDance’s broader approach to capture market share in China’s competitive AI landscape, where over 200 large language models are currently competing for adoption 1.

The In-Depth Research feature’s free testing period is another effort to make advanced AI capabilities accessible and build user adoption, as evidenced by Doubao’s rapid growth to over 26 million monthly active users 1.

ByteDance is positioning itself against domestic players like Baidu and Alibaba while forming alliances with major tech firms to expand its AI ecosystem 1.

2️⃣ Advanced AI capabilities reflect evolution toward specialized use cases

Doubao’s new research feature illustrates how AI platforms are evolving from general-purpose tools to specialized solutions addressing specific complex tasks.

The In-Depth Research function targets high-value use cases including travel planning, purchasing decisions, and policy interpretation—applications that require deep information synthesis rather than simple question answering.

This specialization trend is consistent with broader industry developments, as seen in Doubao’s family of models designed for specific applications like audio recognition and virtual character creation 1.

The feature’s ability to process vast amounts of information and generate comprehensive reports addresses a key pain point for users navigating information overload, reflecting the industry trend toward more sophisticated AI applications.

Recent Doubao developments