Volc Engine Advances AI Capabilities with New Dabao Models

At the FORCE Link AI Innovation Tour in Xiamen on July 30, ByteDance's Volc Engine unveiled significant upgrades to its Dabao AI model series. The announcement included three major innovations: Dabao Image Editing Model 3.0, Simultaneous Interpretation Model 2.0, and the enhanced Dabao Large Model 1.6 series.

Figure: Zhang Tan, President of ByteDance Engine, announced the latest Doubao models

Enhanced AI Tools for Enterprise Applications

The Dabao Image Editing Model 3.0 (SeedEdit3.0) addresses common pain points in AI-assisted editing with improved instruction comprehension and generation quality. Users can now perform complex edits like element removal, lighting adjustment, and component replacement through natural language commands.

The Simultaneous Interpretation Model 2.0 represents a breakthrough in real-time translation technology. By adopting a full-duplex framework, it reduces speech delay from 8-10 seconds to just 2-3 seconds while supporting zero-sample voice replication for authentic accent matching.

Cloud-Native Services for Rapid Deployment

Volc Engine simultaneously announced optimizations to its full-stack AI cloud-native services:

• Open-sourced core capabilities of Coze development platform under Apache 2.0 license
• New self-hosted model solution through Volcano Ark model unit
• Upgraded API system with Responses API featuring native context management

"Our latest innovations provide enterprises with complete solutions from foundational models to development tools," stated Zhang Tan during the presentation.

Key Points:

• Image Editing 3.0 enables complex edits via natural language commands
• Simultaneous Interpretation 2.0 reduces latency by 75% with authentic voice replication
• Large Model 1.6 series offers industry-leading TPOT of 10ms at reduced costs
• Coze platform open-sourced with over 10,000 GitHub stars in three days
• Enterprise solutions now support customized model deployment with elastic computing