Google Deploys Gemini 3.8 Flash in AI Mode in Google Search
Google announced on September 2, 2026 that Gemini 3.8 Flash is now available in AI Mode in Google Search, replacing the Gemini 3.7 Flash model that was deployed there three weeks earlier. The update arrives as part of Google's third Flash model release in six weeks.
What was announced
Google published a blog post introducing two variants of Gemini 3.8: Gemini 3.8 Flash, described as its most intelligent general-purpose model, and Gemini 3.8 Flash Cyber, a specialized cybersecurity variant. The consumer availability section of the announcement states that "3.8 Flash is available to Google AI Pro and Ultra subscribers across the Gemini app, AI Mode in Google Search and Gemini in Google Sheets."
The post was authored by Tulsee Doshi, Senior Director of Product Management, and Raluca Ada Popa, Gemini Security Lead at Google DeepMind.
What changed
AI Mode in Google Search now runs on Gemini 3.8 Flash instead of Gemini 3.7 Flash. Google says the new model delivers "significant improvements from 3.7 Flash across software engineering, agentic tasks, and critical, multi-step reasoning in specialized domains." The pricing for API access remains unchanged at $0.75 per million input tokens and $3.75 per million output tokens.
On the HLE-Verified benchmark, which tests multi-step reasoning across STEM, humanities, and professional fields, Gemini 3.8 Flash scored 54.9%. Google says the model also outperforms 3.7 Flash on the DeepSWE v1.1 long-horizon software engineering benchmark and on domain-specific evaluations including Vals Finance Agent V2 and Harvey's Legal Agent Benchmark.
Details and availability
The deployment in AI Mode is available to Google AI Pro and Ultra subscribers. The announcement does not specify whether free-tier Search users also receive responses generated by 3.8 Flash, or whether the model upgrade applies globally from launch.
Google notes that 3.8 Flash "works harder" on complex tasks, executing extra reasoning steps and calling tools iteratively. The company says the model may use more tokens to maximize performance, particularly at higher effort levels. Developers who need compute efficiency can use lower effort levels or continue with Gemini 3.7 Flash, which remains supported.
Context
Gemini 3.7 Flash was introduced in AI Mode in Google Search in mid-August 2026. The rapid succession of Flash model updates — three in six weeks — indicates Google is iterating quickly on the models that power its AI-generated search responses. The model improvements focus on reasoning depth and coding ability, capabilities that directly affect how AI Mode handles complex queries and multi-step information retrieval.
What has not been confirmed
The announcement does not describe any changes to how AI Mode selects sources, ranks results, or displays citations. Google did not provide information about whether the 3.8 Flash deployment affects all query types equally or only specific categories. There is no published data on how the model change affects traffic to websites cited in AI Mode responses.
Sources
- Introducing Gemini 3.8 Flash and 3.8 Flash Cyber — Google Blog, September 2, 2026
