On June 23, at the 2026 Volcano Engine Force Conference, Doubao's large model officially unveiled its latest flagship version - Doubao-Seed-2.1 Pro. This upgrade marks a critical step forward for domestic large models in core performance and engineering applications.

According to benchmark test data released on-site, this new version demonstrates strong performance capabilities. It has shown the ability to compete with globally top models such as GPT-5.5, Claude Opus 4.7, and Gemini 3.1 Pro in multiple authoritative benchmark tests. This is not just about parameter scale, but also represents a significant leap in the logical reasoning and complex task processing capabilities of domestic AI.

Tan Dai, President of Volcano Engine, emphasized the "production-level transformation" brought by this version. Compared to earlier versions, Doubao-Seed-2.1 Pro has achieved breakthroughs in four core dimensions: code delivery, long-term Agent tasks, multimodal understanding, and enterprise-level stable operations. This means it can not only "understand" complex business requirements, but also has excellent engineering delivery and long-term planning capabilities, effectively serving high-demand enterprise production environments.

At the same time, the popularization process of large models is also accelerating. Just before the conference, data showed that the daily token call volume of the Doubao large model had reached as high as 180 trillion. Such a massive call scale reflects the high user stickiness of this model among developers and enterprise users, gradually moving from the "experimental field" to the front end of various industries' businesses.

From comprehensive parameter capability catching up, to deep accumulation of enterprise-level delivery capabilities, the release of Doubao 2.1 Pro not only demonstrates the resilience of leading domestic large model manufacturers in the technical marathon, but also indicates that AI applications are entering a new development cycle centered around "stability, reliability, and in-depth scenario application." With the official launch of this flagship version, future domestic enterprises will have more high-quality technological infrastructure options when building intelligent production systems.