July 17, 2026 has been circled on developer calendars for weeks. It is the date multiple outlets report Google will finally ship Gemini 3.5 Pro, the reasoning-focused flagship that was supposed to arrive back in June.
What's Actually Confirmed
Google itself has not confirmed a launch, a price, or a spec sheet. What is solidly confirmed is the tier below it: Gemini 3.5 Flash has been generally available since May 2026, priced at $1.50 per million input tokens and $9.00 per million output tokens, with a 1,048,576 token context window.
What is confirmed indirectly is a messier story. Multiple outlets report that Google DeepMind scrapped the original Gemini 3.5 Pro base model and rebuilt it from scratch rather than ship something that underperformed internal targets. This is a far more expensive decision than a routine schedule slip — rebuilding a frontier base model means re-running large portions of pre-training and alignment work measured in months.
The Researcher Exodus
The detail that turned this from a product story into a talent story broke in the same week Google missed its original deadline. Four senior Gemini researchers announced moves to Anthropic between June 21 and June 27, 2026. The timing — in the same seven-day window as the missed deadline — is what gave the story legs beyond the usual industry churn. Losing four senior people with direct knowledge of an in-progress model's architecture, in the exact week that model's release slipped again, invites inevitable speculation about morale.
The Competitive Landscape
Google is not just racing OpenAI and Anthropic on capability anymore. It is racing a cost floor that keeps dropping, largely because of open-weight labs it does not control. A CNBC investigation found that Chinese-built AI models now account for between 30% and 46% of all enterprise API token traffic flowing through US developer platforms, up from just 4.5% in early 2025.
DeepSeek remains the single largest vendor on OpenRouter by token share. Z.ai's GLM-5.2 saw the fastest adoption of any model tracked by Vercel in 2026, with daily token volume growing roughly 27 times over in its first week.
What's Rumoured But Not Confirmed
Third-party estimates, none sourced to an official Google pricing page, put Gemini 3.5 Pro's context window at roughly 2 million tokens — double Flash's — and pricing around $15 per million input tokens and $60 per million output tokens. Reports also describe a "Deep Think" reasoning layer and expanded autonomous workflow capabilities.
Every one of those figures should be read as a placeholder, not a spec. Google has not published a model card, pricing page, or benchmark table for Gemini 3.5 Pro as of this writing.
