Google releases three new Gemini models — but no 3.5 Pro

My take: For teams working with AI in production, Gemini 3.6 Flash is the most relevant piece of this release: $7.50 per million output tokens (lower than its predecessor), 304 tokens per second (roughly four times the median for reasoning models in its price tier), and 17% fewer tokens generated for the same work. Less cost, more speed, more efficiency.

The 1-million-token context window and built-in computer use position it as a serious option for workflows that require long-document analysis or screen-based task automation. The knowledge cutoff has been updated to March 2026.

What's also notable is what didn't arrive: Gemini 3.5 Pro, Google's flagship model, is still not available to the general public. What is clear is that Google is competing first on price and speed in the mid-tier market. Could your current workflow benefit from migrating to a model with these metrics?