Google delays Gemini 3.5 Pro to July

Delaying a major Pro model by a month suggests Google is prioritizing quality and reliability over rushing a launch in the highly competitive agentic space. This cautiousness is likely a response to developer feedback on token efficiency and the real-world utility of its agent frameworks.

• –Google is reportedly focusing on refining long-horizon tasks and agent-based workflows, indicating that Gemini 3.5 Pro is being built with native tool-use and autonomy as core priorities.
• –Feedback from Gemini 3.5 Flash, particularly regarding rapid token consumption, is being integrated to ensure the Pro version is more resource-efficient for developers.
• –Rushing a "Pro" model with agentic bugs would damage developer confidence; taking an extra month to iron out issues is a smart play in a market increasingly dominated by reliable alternatives.
• –The delay occurs amidst high-profile researcher departures from Google, adding pressure to deliver a model that can stand toe-to-toe with Anthropic and OpenAI's latest offerings.