(Latest Update: Google has since removed references to Gemini 2.0 Pro Experimental from its changelog, calling it an “out-of-date release note published in error.” Original story below.)
Google has discreetly announced its next flagship Gemini 2.0 Pro Experimental AI model, opting for an unorthodox launch method: a changelog update for its Gemini chatbot app. This successor to the Gemini 1.5 Pro model (released February 2023) arrives as global attention shifts to Chinese AI startup DeepSeek, whose open-source models rival U.S. tech giants’ offerings.
Gemini 2.0 Pro Experimental, now available to Gemini Advanced subscribers, promises “better factuality” and “stronger performance” in coding, mathematics, and complex tasks like statistical modeling and quantum algorithms. According to Google’s changelog (now revised), the model is in “early preview” and may exhibit “unexpected behaviors” or errors. Unlike other Gemini models, it lacks real-time data access and certain app integrations.

Image Credits:Google
Key Features of Gemini 2.0 Pro Experimental
- Enhanced Coding & Math: Supports generating programs from scratch and solving advanced equations.
- Early Access: Exclusively for Gemini Advanced (Google One AI Premium or Workspace users).
- Experimental Status: Feedback-driven improvements prioritized for future updates.
Competitive Landscape: DeepSeek’s Challenge
The launch coincides with rising competition from DeepSeek, whose open-source AI models reportedly match or exceed U.S. counterparts like GPT-4 and Claude 3. This has sparked debates in Silicon Valley and Washington about AI dominance. Learn more about DeepSeek’s latest models.
Gemini 2.0 Flash Rolls Out Broadly
Alongside Gemini 2.0 Pro, Google has made Gemini 2.0 Flash — its lightweight, cost-efficient model announced in December — the default for all Gemini app users.
Limitations & Future Plans
Google warns that Gemini 2.0 Pro is not yet production-ready. Subscribers are urged to test its capabilities and report issues to refine future iterations. The company emphasizes “rapid iteration” to stay ahead in the global AI race.