Gemini 3.8 Flash Lands in Google Search AI Mode

Free Instant SEO Audit

By NewSunSEO October 9, 2026

Updated October 26, 2026 at 2:25 AM ET

Google AI Pro and Ultra subscribers now have a stronger model behind complex questions in Google Search’s AI Mode: Gemini 3.8 Flash. Robby Stein, VP of Product at Google Search, said on X that the model had "landed in Search" on September 2, 2026, with gains over 3.7 Flash in software engineering, agentic tasks and multi-step reasoning. The rollout is global for subscribers, while free users stay on an earlier model for now.

What changed inside Google AI Mode

Gemini 3.8 Flash is now the selectable model inside AI Mode for Google AI Pro and Ultra subscribers worldwide. Subscribers reach it by tapping the plus icon and choosing the model from the dropdown menu. Stein described it as "our most intelligent workhorse model yet, delivering significant improvements from 3.7 Flash across software engineering, agentic tasks, and multi-step reasoning."

Nick Fox, a head of Google Search, said on X: "Less than three weeks after 3.7, we’re introducing Gemini 3.8 Flash." Google describes Gemini 3.8 Flash as its "most intelligent workhorse model, delivering significant improvements from 3.7 Flash across software engineering, agentic tasks, and critical, multi-step reasoning in specialized domains."

How 3.8 Flash performs on hard tasks

Google reported on its blog that on the DeepSWE v1.1 long-horizon software engineering benchmark, "3.8 Flash outperforms most larger frontier models in autonomously solving complex engineering problems end to end, only at a fraction of the cost." The same page states the model reaches 54.9% on HLE-Verified, a test of multi-step reasoning across STEM, humanities and professional fields, and that it scores ahead of 3.7 Flash on the Vals Finance Agent V2 and Harvey’s Legal Agent benchmarks.

Trying a range of questions can help when you evaluate a model.

Who gets access, and who waits

Access is tied to a paid tier. Google AI Pro and Ultra subscribers can select 3.8 Flash in AI Mode now, globally. Free users do not have it yet, even after 3.7 Flash shipped, and Stein said Google is "looking to bring these capabilities to more people as we can." That leaves two different AI Mode experiences running at once: subscribers on 3.8 Flash and everyone else on an earlier model.

Why the release pace matters for the sites you manage

Shipping 3.8 Flash less than three weeks after 3.7 Flash puts AI Mode on a short revision cycle. It is worth keeping an eye on how client pages perform over time. Checking AI Mode output periodically rather than only once is a sensible habit.

Whether an AI agent can actually read and use a site is what AI Agent Readiness measures, and SEOScanPro tracks AI visibility and how a site shows up to AI search.

What to do now

Review the AI Mode answers your clients’ core queries now return.
Confirm each client site can be read and used by an AI agent.
Track AI visibility on a recurring basis.
Separate subscriber and free experiences when you report to clients: Google AI Pro and Ultra subscribers can select 3.8 Flash, while it is not yet available to free users.

FAQ

What is Gemini 3.8 Flash?

Gemini 3.8 Flash is a Google model available in Google Search’s AI Mode. Robby Stein, VP of Product at Google Search, described it as Google’s most intelligent workhorse model yet, with improvements over 3.7 Flash in software engineering, agentic tasks and multi-step reasoning in specialized domains.

Who can use Gemini 3.8 Flash in Google Search AI Mode?

Google AI Pro and Ultra subscribers can use Gemini 3.8 Flash in AI Mode globally. It is reached by tapping the plus icon and selecting the model from the dropdown menu. Free users do not yet have access.

How is Gemini 3.8 Flash different from 3.7 Flash?

Google says 3.8 Flash delivers significant improvements over 3.7 Flash in software engineering, agentic tasks and multi-step reasoning in specialized domains. The two releases are less than three weeks apart, and Google reported on its blog that on the DeepSWE v1.1 benchmark 3.8 Flash outperforms most larger frontier models in autonomously solving complex engineering problems end to end, at a fraction of the cost.