Google has released three new AI models, including Gemini 3.6 Flash as a direct replacement for Gemini 3.5 Flash, which has already been deprecated2. The flagship Gemini 3.5 Pro, which was supposed to launch in June, remains in testing1.
Gemini 3.6 Flash is positioned as a faster, cheaper model that Google describes as marginally more capable and better at coding than its predecessor, with multimodal features. Google said the changes to 3.6 Flash were made in response to user feedback on the 3.5 release. Gemini 3.5 Flash had drawn criticism for not living up to Google's promises around code generation, according to Ars Technica, which noted that this may have been a consequence of Google's focus on efficiency as businesses have grown concerned over the cost of AI tokens.
Among the three new models announced, one is Google's first version of Gemini geared toward cybersecurity.
The rapid deprecation of Gemini 3.5 Flash is notable given that it was announced at Google I/O in May. ANALYSIS A model lifecycle of roughly two months from launch to deprecation represents a rapid iteration cadence for Google's Gemini line.
Both Business Insider and Ars Technica confirmed that Google says Gemini 3.5 Pro is not yet ready. ◆ The gap between the Pro and Flash release timelines suggests Google is encountering distinct challenges at the higher-capability tier that its efficiency-focused Flash line does not face.
Google's broader strategy, as framed by Business Insider, centers on doubling down on cheaper, faster AI. ◆ The pattern of shipping incremental Flash updates while withholding the Pro model is consistent with that framing — prioritizing cost-efficient inference over frontier capability.
Google is simultaneously navigating legal pressure on its AI training practices. Publishers including Hachette and author Scott Turow filed a copyright infringement lawsuit against Google in federal court in New York, accusing the company of illegally using millions of copyrighted books to train its Gemini models ctx.