Gemini 3.5 Pro: June 2026 Release, Leaks and What Businesses Can Expect After Flash & Omni

On May 19, 2026, Google unveiled the next generation of its AI models at its annual I/O developer conference. Alongside the already available Gemini 3.5 Flash and the multimodal Gemini Omni, CEO Sundar Pichai announced Gemini 3.5 Pro for June 2026. While Flash focuses on speed and agentic capabilities, Pro is set to elevate pure reasoning and coding performance to a new level.

This article summarizes what is verifiably known after the keynote - and what conclusions businesses can draw if they strategically deploy Google's AI stack.

What Google Actually Announced at I/O 2026

I/O 2026 was the biggest AI-focused event since the launch of the Gemini family. Sundar Pichai shared three key data points:

• Gemini 3.5 Flash is available immediately (May 19, 2026) in the Gemini app, AI Mode in Google Search, and for developers in Google AI Studio, Android Studio, and via the Gemini API (SiliconANGLE, May 2026).
• Gemini 3.5 Pro will be released in June 2026 according to Pichai: "Give us until next month to get it to you" (Wavespeed AI, May 2026).
• Gemini Omni Flash, Google's new multimodal video model, is also available from May 19, 2026 for Google AI Plus, Pro, and Ultra subscribers (Google Blog, May 2026).

Important: Neither Gemini 3.5 Pro nor a later Omni Pro version currently have official benchmarks, pricing, or API IDs. All technical details about Pro below are derived from Flash's behavior or from official Google statements.

Gemini 3.5 Flash: The Preview of Pro

Flash is the smallest and fastest model in the 3.5 series - yet according to Google it surpasses Gemini 3.1 Pro, released just four months ago, on most benchmarks. This is remarkable because it means Pro must not only beat 3.1 Pro but significantly outperform Flash - otherwise the differentiation would not hold up.

Concrete numbers for Gemini 3.5 Flash:

Flash achieves better scores than Gemini 3.1 Pro on Terminal-Bench 2.1, GDPval-AA, and MCP Atlas. At the same time, there is a notable regression on specific reasoning and long-context benchmarks - exactly where Pro will need to score later (Wavespeed AI, May 2026).

What We Know About Gemini 3.5 Pro - and What We Don't

Confirmed

• Release timeframe: June 2026, according to Sundar Pichai's direct statement on the I/O stage.
• Focus: Coding, agentic workflows, and complex reasoning - the same emphasis as Flash but at a higher level.
• Internal test status: Pichai confirmed that 3.5 Pro is already running internally at Google.

Not confirmed (speculation)

• Concrete benchmarks: Google has not published any numbers.
• Pricing structure: Likely higher than Flash but cheaper than 3.1 Pro was at its release.
• Context window: Likely 1 million tokens, possibly more.
• API ID: No model card or documentation published yet.

The delay of Pro relative to Flash is typical of Google's release pattern: Flash is rolled out as a fast, broadly available model, while Pro undergoes additional safety testing and fine-tuning. The one-month window suggests a relatively short polish cycle - which in turn indicates that 3.5 Pro is not an entirely new model but a scaled version of the same architecture stack as Flash.

Gemini Omni: Why the Video Model Matters for Pro

Alongside the text-based models, Google introduced Gemini Omni, a new model family for multimodal content creation. Omni Flash can generate new videos from images, audio, video, and text - and edit them conversationally afterwards (Google Blog, May 2026).

This is relevant for businesses because Google is signaling that the 3.5 architecture powers not only text-based agents but also serves as the "reasoning engine" behind Omni. When 3.5 Pro launches, integration into Omni Pro is likely - with significantly higher video quality and longer sequences.

Gemini Spark: The 24/7 Agent Running in the Background

Another I/O announcement that is often overlooked: Gemini Spark. Spark is a persistent cloud agent running on dedicated virtual machines in Google Cloud, remaining active even when the user's device is turned off (TechSphere News, May 2026).

Spark uses Gemini 3.5 as its foundation and is planned for a beta release for Ultra subscribers by the end of 2026. For businesses this means: the combination of 3.5 Pro and Spark could enable long-running, autonomous business processes - from continuous market monitoring to fully automated reporting.

What Companies Should Do Now

1. Test Flash, Don't Overlook Pro

Companies not yet using Gemini productively should start with 3.5 Flash. The low cost ($1.50/$9.00 per MTok) and high speed make Flash ideal for internal tests. At the same time, developers should prepare their SDKs and configurations so that switching to 3.5 Pro in June is just a parameter change.

2. Prioritize Agentic Workflows

The entire 3.5 series is explicitly built for agentic use - not just chat. Chief AI Architect Koray Kavukcuoglu emphasized at I/O that 3.5 Flash can sustain autonomous sessions for several hours and execute complex coding pipelines independently (TechSphere News, May 2026).

Businesses should therefore not ask "How do we replace ChatGPT?" but rather "Which processes can we automate with an agent that works autonomously for hours?"

3. Plan a Multi-Model Architecture

The best AI infrastructure in 2026 is not a monolith but a routed stack:

4. Consider SynthID and Compliance

All videos generated with Omni carry Google's invisible SynthID watermark. Companies using AI-generated media in marketing or internal training should adjust their compliance policies accordingly - particularly with regard to the EU AI Act, which from August 2026 mandates transparency obligations for AI-generated content.

FAQ: Common Questions About Gemini 3.5 Pro

When will Gemini 3.5 Pro be released? Officially in June 2026. Sundar Pichai said at I/O 2026 on May 19: "Give us until next month to get it to you."

Is Gemini 3.5 Pro already available? No. As of May 29, 2026, there is neither a public API nor a model card. The currently available 3.5 model is Flash.

What distinguishes 3.5 Pro from 3.5 Flash? Flash is optimized for maximum speed and cost efficiency and already surpasses 3.1 Pro on many benchmarks. Pro will be Google's full flagship model for complex reasoning, coding, and long contexts.

How much will Gemini 3.5 Pro cost? Google has not announced pricing. As a reference: Flash costs $1.50 input / $9.00 output per million tokens. Pro will likely be more expensive but below the old 3.1 Pro prices.

What is Gemini Omni? A new multimodal model for video creation and editing from any input (image, audio, video, text). Omni Flash has been available since May 19, 2026; Omni Pro will follow later.

Should my company wait for 3.5 Pro? No. 3.5 Flash is already production-ready and allows immediate testing. Companies that build their architecture in a parameterized way can switch to Pro within minutes in June.

What does Gemini 3.5 mean for competition with OpenAI and Anthropic? With Flash, Google has released a model that is faster and cheaper than comparable models from OpenAI and Anthropic. Whether Pro can maintain this lead on reasoning benchmarks will be the decisive question of summer 2026.

Conclusion: A Strategically Important Summer for Google

Gemini 3.5 Pro is more than just a model update - it is Google's response to the rapid development at OpenAI (GPT-5.5) and Anthropic (Claude Opus 4.8). That Flash already surpasses 3.1 Pro sets the bar high. If Pro delivers the expected leaps in reasoning and coding in June, Google could cement its claim as the leading enterprise AI platform.

For businesses, the pragmatic path is clear: test with Flash now, build agentic workflows, and parameterize infrastructure so Pro can be integrated seamlessly in June. Those who wait for June without experimenting with Flash first lose valuable time - and leave the competitive advantage of the next AI generation on the table.

Not sure which AI model is right for your processes? Schedule a free initial consultation - we analyze your workflows and recommend the optimal model stack.

---

Sources

• Google Blog (19.05.2026): Introducing Gemini Omni
• SiliconANGLE (19.05.2026): Google targets AI agents and video generation with Gemini 3.5 Flash and Omni
• Wavespeed AI (20.05.2026): Gemini 3.5 Pro Is Coming Next Month - What the Flash Release Already Tells Us
• Gadgets 360 (20.05.2026): Google IO 2026: Here's Everything That Was Announced
• TechSphere News (19.05.2026): Google I/O 2025: Gemini 3.5 Flash, Omni Multimodal Model, and 24/7 Cloud Agent Spark
• The New Stack (19.05.2026): Google's Gemini 3.5 Flash beats the frontier models
• ScriptByAI (20.05.2026): Gemini & Google AI Timeline
• Fello AI (20.05.2026): All We Know About Google Gemini 4: Release Date, Rumors & What to Expect