Google launches Gemini 3.5 Flash and Gemini Omni at Google I/O 2026

Google launched Gemini 3.5 Flash on May 19, 2026 at Google I/O 2026, the first model in a new series it describes as combining frontier-level intelligence with Flash-tier speed. Alongside it, Google announced Gemini Omni, a multimodal video-generation model that accepts images, audio, video, and text as inputs and allows users to edit video through natural-language conversation.

What's new

Gemini 3.5 Flash

According to Google, "Gemini 3.5 Flash delivers intelligence that rivals large flagship models on multiple dimensions, at the speeds you have come to expect from the Flash series." Key specifics:

• Outperforms Gemini 3.1 Pro on coding and agentic benchmarks at flash speeds
• Claims to process output tokens roughly 4x faster than other frontier models
• Benchmark performance: Terminal-Bench 2.1 (76.2%), MCP Atlas (83.6%), CharXiv Reasoning — multimodal (84.2%)
• Priced at under half the cost of other frontier models, per Google
• Powers Gemini Spark, Google's personal AI agent product (beta for Google AI Ultra subscribers in the US)
• Available at launch on: Gemini API, Google AI Studio, Android Studio, Gemini app (global), AI Mode in Google Search, and Gemini Enterprise Agent Platform

Gemini 3.5 Pro is already in internal use at Google; a broader rollout was expected in June 2026.

Gemini Omni

Gemini Omni is a multimodal model that accepts any combination of image, audio, video, and text inputs to generate and edit video. Users can edit videos conversationally — changing environments, adding objects, adjusting camera angles — while the model maintains character consistency, physics, and scene continuity across edits.

Gemini Omni began rolling out May 29, 2026 to Google AI Plus, Pro, and Ultra subscribers globally via the Gemini app and Google Flow. API access for developers and enterprise customers was announced for the following weeks. YouTube Shorts and YouTube Create users will also gain access at no cost.

Context

Google I/O 2026 was framed around what Google called the "agentic Gemini era" — AI that doesn't just answer questions but takes multi-step actions across products. Gemini 3.5 Flash powers new interactive-UI features in Search, Gemini Workspace integrations, and the Gemini Spark personal agent.

The timing marks a shift in Google's competitive positioning. Gemini 3.1 Pro had been the frontier reference; 3.5 Flash is designed to make that performance level available at speed-and-cost points suitable for production agent deployments — competing with Anthropic's Claude Opus 4.8 and OpenAI's GPT-5.5 family for enterprise agentic workloads.

Why it matters

For developers building agentic applications, the cost/performance positioning of Gemini 3.5 Flash is significant. If the benchmarks hold in production, teams could run complex multi-step workflows at frontier quality without the premium pricing of larger models. The 4x speed claim matters for latency-sensitive applications — real-time agents, voice pipelines, interactive UI generation.

Gemini Omni addresses a capability gap no other frontier lab has filled in the same way: conversational video editing with maintained scene coherence. Whether this extends beyond consumer use (creative apps, YouTube) into enterprise workflows (product demos, training videos) will determine its practical impact.