The verdict: Gemini 3.5 Pro has not reached general availability by June 30, 2026. Google originally announced a June 2026 target at Google I/O on May 19. As of this morning, the model remains in limited Vertex AI enterprise preview for select partners only. Google has now confirmed a July 2026 general availability release.
What we have been tracking and why it matters
Gemini 3.5 Pro has been the most-watched model deadline of Q2 2026 for three reasons. First, its promised specifications — a 2-million-token context window, a Deep Think reasoning mode, and full Google Workspace integration — are meaningfully different from what is currently available. Second, Google's Workspace integration would make it natively accessible to UK businesses already using Gmail, Docs, and Drive, reducing the barrier to adoption. Third, the June 30 deadline created a clear decision point for UK businesses weighing platform choices.
Prediction markets had the June 30 release at roughly 50–55% probability. The miss closes the Q2 chapter and moves the decision to July.
What is available now, while you wait
The June 30 miss does not mean Google has nothing to offer today. Three Gemini models are already publicly available:
- Gemini 3.5 Flash — lighter and faster than Pro, already in Google AI Studio and the Gemini API. Handles routine summarisation, drafting, and Q&A tasks well. Free tier available.
- Gemini for Google Workspace — integrated into Gmail, Docs, Sheets, and Slides. If you use Google Workspace and have not enabled Gemini, this is the easiest starting point. Available on Business Starter and above.
- Gemini Advanced — the consumer-facing tier at the top of Google One, using the current Gemini 1.5 Ultra base. Useful for complex reasoning and longer documents, though not at the 2M token level Pro will offer.
For most UK small business tasks — writing emails, summarising documents, drafting quotes, answering customer questions — Gemini 3.5 Flash and the Workspace integration are adequate today. The Pro model is worth waiting for specifically if you need extremely large context windows or the Deep Think reasoning mode for complex multi-step analysis.
What to expect when Pro actually ships in July
At 2 million tokens, you can load approximately 1,500 pages of dense text into a single conversation. For UK service businesses, the practical uses include: processing a full year of email threads with a customer, loading multiple supplier contracts simultaneously, analysing complete project files without chunking. This is the headline feature that makes Pro materially different from Flash.
Deep Think is designed for multi-step problems that require planning rather than pattern matching — complex financial analysis, detailed project scheduling, strategic planning with multiple constraints. For most routine UK business tasks it will not be the default mode, but for occasional complex decisions it could be the most practical AI reasoning tool available through a Google Workspace subscription.
The integration planned for Pro goes beyond what is currently available in Workspace Gemini — it includes deeper access to your Drive files, Calendar context, and Gmail history within a single session. For businesses already operating primarily in Google Workspace, this removes the need to copy and paste between tools.
The deadline miss in context
Model release delays are normal in the current competitive environment. OpenAI has delayed several releases in 2026. Anthropic extended testing windows before shipping Claude Fable 5. Google's delay on Gemini 3.5 Pro is likely cautious testing rather than a fundamental technical problem — the model has been in Vertex AI enterprise preview since late June, which means it is substantially complete.
The more relevant question for UK businesses is not whether Google missed a self-imposed date but whether the model, when it ships in July, does what it promises. That is a question to answer with a test case in hand.
Operator action: three decisions for July
Decision 1 — Prepare your test case now. Before Gemini 3.5 Pro ships, write down one specific task that would benefit from a 2-million-token context window or advanced reasoning. Your evaluation will be faster and more useful if you have a benchmark task ready rather than doing a generic "is it better?" test when it lands.
Decision 2 — Start with Flash today. If you have not used any Gemini product, start with Gemini 3.5 Flash in Google AI Studio or the Workspace Gemini assistant this week. Understanding how Gemini approaches tasks will make your Pro evaluation more meaningful when it arrives.
Decision 3 — Do not extend contracts waiting for Pro. If you are evaluating other AI tools and delaying decisions to see what Gemini Pro delivers, stop waiting. Evaluate what is available now. Pro's July arrival may or may not change your calculus — test it when it ships rather than waiting in limbo.