ByteDance Unveils Four Advanced AI Models with Enhanced Features
ByteDance Unveils Four Advanced AI Models with Enhanced Features
At the recent FORCE LINK AI Innovation Tour in Wuhan, ByteDance's Volcano Engine officially introduced four new or upgraded Doubao large models. The release includes the upgraded Doubao Large Model 1.6, the lightweight Doubao Large Model 1.6lite, and two newly launched models: Doubao Voice Synthesis Model 2.0 and Doubao Voice Cloning Model 2.0. This upgrade aims to enhance intelligence levels and provide enterprises with more flexible AI solutions.
Enhanced Performance and Flexibility
The upgraded Doubao Large Model 1.6 stands out as a key highlight of this release. It natively supports four thinking lengths: Minimal, Low, Medium, and High, allowing users to tailor the model's response effect, latency, and cost to their specific needs. This innovation ensures optimal performance across diverse scenarios. Compared to its predecessor, the upgraded model reduces output tokens by 77.5% and cuts thinking time by 84.6%, all while maintaining consistent performance.

Lightweight Alternative for Cost-Efficiency
To address the demand for high-performance yet cost-effective solutions, ByteDance also introduced the Doubao Large Model 1.6lite. This lighter version offers faster inference speeds and better value for money. In enterprise-level tests, 1.6lite outperformed Doubao Large Model 1.5pro by 14%. Notably, in the frequently used input range of 0-32k, overall usage costs dropped by 53.3%.
Breakthroughs in Voice Synthesis
Beyond large model upgrades, ByteDance unveiled two advanced voice models:
- Doubao Voice Synthesis Model 2.0: Boasts stronger emotional expression and instruction-following capabilities, enabling accurate reading of complex formulas.
- Doubao Voice Cloning Model 2.0: Leverages deep semantic understanding powered by the Doubao large language model for precise emotional expression.
Users can now adjust speed, emotion, and tone through natural language commands, resulting in highly personalized voice outputs.
Key Points
- ByteDance launched four new or upgraded AI models under its Volcano Engine.
- The upgraded Doubao Large Model 1.6 reduces output tokens by 77.5% and thinking time by 84.6%.
- The lightweight Doubao Large Model 1.6lite cuts costs by 53.3% while improving performance.
- New voice synthesis models offer enhanced emotional expression and personalization.
- These innovations demonstrate ByteDance's continued leadership in AI development.