At Google I/O on May 20, 2026, Google introduced Gemini 3.5 Flash, the multimodal Gemini Omni family, and a suite of agent development and Workspace integrations designed to speed enterprise adoption of agentic and multimodal applications;
At Google I/O on May 20, 2026, Google unveiled Gemini 3.5 Flash, the Gemini Omni family and an expanded set of agent development tools aimed at helping enterprises move agentic and multimodal prototypes into production. The release bundles a low-latency model optimized for long-horizon agentic and coding tasks with platform features and Workspace intelligence that Google says will accelerate internal developer teams and business workflows.
Gemini 3.5 Flash, the first model in the 3.5 family, is positioned for sustained agentic reasoning and coding work with a claimed balance of speed and frontier performance. Google published benchmark figures including Terminal — Bench 2.1: 76.2%, GDPval — AA: 1656 Elo, MCP Atlas: 83.6% and multimodal CharXiv: 84.2%. The company also said 3.5 Flash often runs at less than half the cost of comparable models; a higher‑capability Gemini 3.5 Pro is in testing and expected next month.
Gemini Omni extends multimodal generation to video, accepting text, audio, image and video inputs to produce and edit dynamic visual content. Google positioned Omni for enterprise scenarios such as interactive virtual try‑ons, complex post‑production workflows and tailored video narratives; Gemini Omni Flash will be made available to developers and enterprise customers via the Gemini API and Agent Platform in the coming weeks.
Platform — level updates emphasize agentic development and secure deployment. Expanded Antigravity capabilities are integrated with the Agent Platform, a Managed Agents API will run custom agents inside secure Google‑hosted environments, and Gemini Spark is described as a 24/7 personal AI agent for Gemini Enterprise and Workspace customers. Workspace additions include Google Pics for image generation and editing plus new voice features rolling into Gmail, Docs and Keep.
The Agent Platform additions target real deployment needs: developers can build agents with Gemini 3.5 Flash in Google AI Studio, on the Gemini Enterprise Agent Platform or via Antigravity, while business users can access capabilities through the Gemini Enterprise app. Security and maintenance tooling was stressed as well — CodeMender, delivered through the Agent Platform, is presented as an AI security agent to find and help fix code vulnerabilities.
Google said these models were engineered by Google DeepMind with purpose‑built infrastructure and a co‑design of model and hardware as part of its Next ’26 blueprint, which includes eighth‑generation TPUs and an Agentic Data Cloud. Key immediate timelines: Gemini 3.5 Flash is rolling out today (May 20, 2026), Gemini Omni Flash is due in the coming weeks, and Gemini 3.5 Pro is slated for release after testing next month.
Sources
Replies (0)
No replies in this topic yet.
