Baidu Launches Token Exchange for Verified AI Model Services

Unite.AI
Read full story· 4 min
Baidu launched Token Exchange, a marketplace for verified AI model services, averaging 50M calls daily after one month. It also released ERNIE 5.5 Flash Preview, a large post-trained model with 280B parameters, plus two application-specific models.

What happened

On November 12, 2026, Baidu introduced Token Exchange, a marketplace where users can compare and purchase verified AI model services, at its Baidu World 2026 event. After one month, the platform averaged over 50 million calls and more than 700 billion tokens per day. Baidu also launched ERNIE 5.5 Flash Preview, a large post-trained model with 280 billion parameters, and released two additional application-specific models. These announcements were part of a broader update to Baidu's AI products, infrastructure, and agent tools.

Key details

  • Token Exchange verifies AI models through fingerprinting, capability tests, and ongoing monitoring.
  • Over 100 model resources are available on Token Exchange, with providers rated by regular checks.
  • ERNIE 5.5 Flash Preview has 280 billion parameters, a 1-million-token context window, and is priced at RMB 1 per million input tokens and RMB 4 per million output tokens.
  • Baidu plans to release ERNIE 5.5 Flash Preview's weights under Apache 2.0 by November 25, 2026.
  • The two new application models are ERNIE Wizshot Micro-Expression for expressive video and ERNIE Light Deep Thinking for complex search and tool use.

Why it matters

Businesses and individual users now have a centralized platform to compare, verify, and access AI model services from multiple providers, potentially reducing the time and risk involved in selecting models. Model providers face ongoing verification and credit ratings, which could affect their visibility and credibility on the marketplace. Developers and enterprises gain access to new, large-scale models and specialized tools, with transparent pricing and upcoming open weights, enabling broader experimentation and deployment.


More in Business & Enterprise