Jakarta, INTI - Technology company ByteDance released its artificial intelligence (AI) model Doubao 2.0 on Saturday, February 14, 2026. This is an upgraded AI model from the previous version of Doubao.
Doubao 2.0 is positioned as entering the “agent era,” meaning it is expected to carry out complex real-world tasks rather than merely answering questions.
Doubao 2.0 comes in two versions: a “regular” version and a Pro version that includes complex reasoning capabilities and the ability to execute multi-step tasks.
The Pro version is claimed to rival the capabilities of GPT-5.1 developed by OpenAI as well as Gemini 3 Pro from Google.
More Affordable
The difference is that Doubao 2.0 Pro is designed to be more affordable than the two AI models from the United States.
“This cost advantage is becoming increasingly important because complex real-world tasks involve larger-scale inference and multi-stage generation (AI processing across multiple steps), which will consume a large number of tokens,” ByteDance said.
Doubao 2.0 reflects ByteDance’s ambition to offer services beyond a simple chatbot, moving toward autonomous agents capable of executing end-to-end tasks.
The AI development team stated that the previous version of Doubao was indeed able to solve problems similar to those found in academic competitions. However, it struggled with tasks such as creating simple programs due to limited ability to independently build efficient workflows.
Doubao 2.0 aims to address these shortcomings by expanding domain knowledge coverage and strengthening instruction-following capabilities, enabling greater consistency and control.
Enhanced Multimodal Capabilities
Multimodal understanding in this AI model has also been significantly improved. This effort was made to meet user needs when processing unstructured materials, such as mixed documents and graphics.
In addition, the model is equipped with enhanced visual reasoning, spatial awareness, and long-context understanding.
Another improvement is the enhanced perception of time series and motion, enabling real-time video analysis.
This feature opens up usage opportunities across various scenarios, such as fitness training, fashion recommendations, and care support.
The Doubao 2.0 AI model is now available on the Doubao app and the web version of Doubao, particularly through “Expert Mode,” as compiled by KompasTekno from Caixin Global.
Conclusion
The launch of Doubao 2.0 marks a major step in ByteDance’s ambition to move beyond conventional chatbots toward autonomous AI agents capable of handling complex real-world tasks. With stronger reasoning, advanced multimodal capabilities, and significantly lower costs, Doubao 2.0 positions itself as a serious competitor to leading AI models from OpenAI and Google.
As enterprises increasingly demand scalable, efficient, and intelligent automation, ByteDance’s latest AI innovation could accelerate the adoption of agent-based systems across industries, from digital services and creative sectors to healthcare, smart cities, and industrial transformation.
Read more: OpenAI Launches GPT-5.3 Codex Spark, a High-Speed and Efficient AI Model for Developers