Gemini 3.5 Pro launch and what it changes for your AI work

Google is keeping everyone guessing on timing. Unveiled at Google I/O on May 19, 2026, Gemini 3.5 Pro has so far reached only a handful of enterprise customers on Vertex AI. On stage, Sundar Pichai asked for a little more time and said Google needed the following month to deliver it, a line that stuck. Everything therefore points to a public release in the second half of June 2026, give or take a few days.

The wait makes sense given what the model promises. Gemini is expected to take a step forward in reasoning and context size, two areas where the competition is moving fast. Before you commit to anything, here is what's actually known about this launch, the confirmed figures and how to make use of it without changing all your tools or piling up subscriptions.

01Key takeaways

• Unveiled on May 19, 2026, Gemini 3.5 Pro should be released to everyone in the second half of June.
• As of June 19, it was still in limited access for enterprise customers on Vertex AI, not yet in the consumer app.
• Announced specs include a context window of 2 million tokens and a Deep Think reasoning mode.
• Expected pricing sits around $15 and $60 per million tokens, compared with $1.50 and $9 for the Flash version already available.
• Anthropic and xAI are also shipping in June 2026, which makes a multi-model tool more useful than ever.

02What Google announced about Gemini 3.5 Pro

The model is positioned as the top tier of the Gemini family. Google highlights a context window of 2 million tokens, enough to take in very long documents or entire codebases in one go. On top of that comes Deep Think, a reasoning mode built for complex problems that need several steps. The Flash version, released on May 19, already serves as the default model in Google Search.

On pricing, the gap is confirmed. Where Flash costs about $1.50 and $9 per million tokens for input and output, the Pro version is expected closer to $15 and $60. That price level keeps it for heavy tasks, where its power makes a difference. For everything else a lighter model is often enough, as the recent price cut showed in consumer AI.

Update of October 16, 2026. Gemini 3.5 Pro never reached general availability. On October 1, 2026, Reuters reported that Google no longer plans to release it and has announced Gemini 4 Argon as its new top model. Google's official API price list now shows Argon at $2 per million input tokens and $10 per million output tokens through November 30, 2026, then $4 and $20. Requests sent to Gemini 3.5 Flash are now served by Gemini 3.6 Flash, billed $0.75 and $3.75 through December 31, 2026, then $1.50 and $7.50.

03Why this launch matters to you

A new top model gives you one more option, with no obligation attached. What counts is less who leads the rankings this month than your ability to test the newcomer on your own tasks. A long contract to analyze, code to review or a case file to summarize are the jobs where a 2 million token context window makes a real difference, as long as the gain justifies the extra cost.

The timing sets the tone. Within a few weeks, Anthropic and xAI are pushing their own releases too, after launches already covered here such as the Fable 5 model and the GPT-5.6 leaks. For users, this race brings prices closer together and pulls quality up. You still need to be able to switch engines without rewriting everything at each release.

04How to use it without piling up subscriptions

The answer comes down to one simple principle, which is to pay per use instead of a flat fee. With your own API key, you pay only for the tokens you consume, with no monthly subscription for each provider. The day Gemini 3.5 Pro opens to everyone, you switch it on in a few seconds and test it on a real case without starting from scratch.

The second habit is to avoid locking yourself in with a single company. A tool that lets you compare several models side by side lets you judge on the evidence, using your real documents. That way you keep the best option of the moment, whether it comes from Google, Anthropic or someone else, then adjust your spending to the task rather than to a yearly contract.

05Frequently asked questions

When does Gemini 3.5 Pro come out?

Google is aiming for a public release in June 2026, with no official date as of June 19. At Google I/O on May 19, Sundar Pichai spoke of the following month. The model was then still in limited access for enterprise customers on Vertex AI, which points to the second half of June for a wider opening.

What's new in Gemini 3.5 Pro?

The two headline features are a context window of 2 million tokens and a reasoning mode called Deep Think. The first lets you process very long documents in one pass. The second targets complex problems that require reasoning in several steps, where a quick single answer falls short.

How much does Gemini 3.5 Pro cost?

The announced price is around $15 per million input tokens and $60 per million output tokens. That's much more than the Flash version, billed at about $1.50 and $9. This price keeps the model for heavy tasks, where its power is worth paying for compared with a lighter model.

Do you need a subscription to use Gemini 3.5 Pro?

Not necessarily. Through an API key you pay only for the tokens you use, with no monthly commitment. A multi-model tool lets you switch Gemini on as soon as it opens and compare it with other models on your own tasks, then adjust your spending to what you really need.

Transparency note. The information above was checked on June 20, 2026 against the Google I/O announcements and the specialist press, with links below. The release date and prices are those announced by Google and may change until the model becomes generally available.

06Sources

1. Google, "Sundar Pichai's opening keynote at Google I/O 2026", May 19, 2026.
2. TechTimes, "Google Gemini 3.5 Pro nears June launch with 2M token context and Deep Think", June 6, 2026.
3. Codersera, the June 2026 launch guide to Gemini 3.5 Pro, 2026.
4. Reuters via Khaleej Times, "Google announces Gemini 4 flagship AI model after months of delays", October 1, 2026.
5. Google, official Gemini API price list, accessed October 16, 2026.
6. Google, Gemini API model deprecations (3.5 Flash routed to 3.6 Flash), accessed October 16, 2026.