Doubao AI Takes a Quantum Leap with Version 2.0

Volcano Engine has unveiled its most significant AI upgrade yet - Doubao Large Model 2.0 delivers dramatic improvements across the board while making the technology more accessible than ever.

Cost Meets Performance

The new version achieves what many thought impossible - better performance at a fraction of the cost. Inference expenses have plummeted by approximately 90% compared to industry leaders, while daily Tokens usage has skyrocketed over 500 times since initial release.

"We've rebuilt the engine while keeping what works," explains a Volcano spokesperson. "The result is AI that doesn't just think better - it thinks cheaper."

Four Models, Endless Possibilities

The update introduces four specialized versions:

• Pro: The powerhouse for complex reasoning and agent tasks
• Lite: Smarter than version 1.8 but lighter on resources
• Mini: Lightning-fast responses at budget-friendly rates
• Code: A developer's dream, optimized for real programming environments

Each model serves distinct needs, from cost-conscious implementations to high-performance applications.

Seeing Is Believing: Multimodal Breakthroughs

The visual understanding capabilities represent some of the most impressive advances:

• Outperforms Gemini3pro in spatial, motion and video comprehension benchmarks
• Excels at interpreting charts and technical diagrams (CharXiv-RQ)
• Surpasses human performance in EgoTempo video analysis tests

The model can now track billiard movements frame-by-frame or provide real-time fitness coaching - capabilities that seemed futuristic just months ago.

Coding Gets Smarter

The Code version brings particularly exciting developments:

• Seamless integration with mainstream IDEs
• Front-end optimizations that slash development time
• Ability to build complex web apps in just five prompt rounds

The "AI Temple Fair" demo showcases this potential - an intricate application now available as open-source material.

Availability and Pricing

The waitlist is gone - APIs are now openly available through Volcano Ark. New developers can start experimenting for as little as ¥8 in their first month.

The Coding Plan package update specifically addresses surging Tokens usage in agent applications, offering smarter resource allocation for programming tasks.

Key Points:

• Inference costs reduced by ~90%
• Four specialized models cover all use cases
• Video understanding now surpasses human benchmarks
• Developer tools significantly enhanced
• APIs immediately available with affordable entry pricing