Google plans next Gemini 2.0 model for December release | The

Google plans to announce its next-generation Gemini 2.0 model in December, according to sources familiar with the plan. The move positions Google for a direct showdown with OpenAI, which is eyeing the same month for its next flagship AI model. Unlike OpenAI's phased rollout with business partners first, Google intends to release Gemini 2.0 widely from the outset. However, insiders report that the model isn't showing the performance gains the Demis Hassabis-led team had hoped for, though interesting new capabilities are still expected. Separately, AI researcher Noam Shazeer—whom Google paid heavily to abandon Character.AI—is working on a reasoning model to rival OpenAI's o1.

The broader AI race is accelerating: Elon Musk's xAI is training Grok 3 on a cluster of 100,000 Nvidia H100s in Memphis, and Meta is training Llama 4 on an even larger chip cluster. Meanwhile, Anthropic quietly released a 'computer use' agent for Claude that can take over a user's screen—albeit buggy and expensive for now. The trend raises questions about whether frontier models are converging in capabilities and commoditizing each other, with real value shifting to the products they power. Despite potential diminishing returns, competitive pressure keeps the race for ever-bigger models intense.