Back to all articles
•Technology

Google DeepMind Unveils New AI Models, Delays Flagship Release Amidst Tough Competition

View original source

Google DeepMind has launched a new suite of AI models including Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber on Tuesday, aiming to enhance efficiency, cost-effectiveness, and cybersecurity capabilities. The models target different segments:

  • Gemini 3.6 Flash: Promises improved performance in coding, knowledge work, and multimodal tasks, reducing token usage by 17% compared to its predecessor.
  • Gemini 3.5 Flash-Lite: Offers the most cost-effective solution in the lineup.
  • Gemini 3.5 Flash Cyber: Specifically designed for cybersecurity, this model will be available only to governments and select partners.

Google emphasizes that these releases aim to deliver efficiency, latency, and reliability for large-scale AI deployment. The update, however, notably lacks the anticipated release of Gemini Pro, Google's high-capability model for complex tasks, delayed due to internal performance challenges. This delay is significant in the context of rapid advancements by rivals like OpenAI, which has launched GPT-5.5 and started GPT-5.6 rollouts, and Anthropic, with its recent releases including Claude Opus 4.8 and Claude Sonnet 5.

The absence of Gemini Pro highlights the competitive pressure Google faces in the high-stakes AI landscape.