Google is moving faster on its budget-friendly Gemini line, but the speed is drawing attention to what has not arrived: the flagship 3.5 Pro model promised for June.
At Google I/O in May, the company said Gemini 3.5 Pro would launch the following month. That deadline passed, leaving developers and businesses waiting for the higher-end release while Google continued updating Flash instead.1
Three weeks after Gemini 3.6 Flash debuted, Google began rolling out Gemini 3.7 Flash as its replacement. The company calls it a “workhorse” shaped by core optimizations and developer feedback, with stronger coding and agentic performance. Google says its FrontierCode score rose from 34.4% to 43.6%, while its AutomationBench result climbed from 17% to 30.4%.1
The commercial pitch is just as aggressive. Through year-end, 3.7 Flash is priced at $0.75 per million input tokens and $3.75 per million output tokens—half the listed cost of 3.6 Flash. Yet the upgrade lands amid cheaper competition and skepticism that incremental benchmark gains justify replacing a model barely a month old.1
Google has also put the new model to work inside Gemini Spark, its AI agent for Workspace tasks. The company says 3.7 Flash makes Spark “more efficient for knowledge work,” improving tool use, accuracy and output quality across complex, multi-skill workflows.2
That deployment exposes the split in Google’s strategy. Subscribers to AI Pro and Ultra can access 3.7 Flash through Spark, while the regular Gemini chatbot still runs on 3.6 Flash. For customers awaiting a true flagship, the latest Flash release looks both like progress and a conspicuous placeholder.