Google added Gemini 3.7 Flash to AI Mode on August 14, 2026, one day after the model’s public release. The update makes the model selectable inside AI Mode’s interface, but only for paid subscribers. For free-tier users, nothing changes yet.
What Is Gemini 3.7 Flash?
Gemini 3.7 Flash is the newest member of Google’s Flash model family, described in Google’s August 13 blog post as “our most intelligent workhorse model yet for coding and agents.” It arrived just three weeks after Gemini 3.6 Flash, built from developer feedback and what Google describes as algorithmic innovations it plans to carry into future models.
According to Google’s announcement, the benchmarks show substantial gains over 3.6 Flash:
- FrontierCode 1.1 Main (a code-generation benchmark): 43.6% vs. 34.4% for 3.6 Flash
- DeepSWE v1.1 (software engineering evaluation): 65.3% vs. 49.0%
- WebDev Arena Elo score: 1,588 vs. 1,538
- GDP.pdf benchmark (a test for processing complex documents): 34.0% vs. 22.0%
- AutomationBench (real-world business workflow automation): 30.4% vs. 17.0%
Google launched the model at an introductory price of $0.75 per million input tokens and $3.75 per million output tokens, which Google says is half the original per-token price of 3.6 Flash. That pricing is available through the end of the year.
How Do You Select It in AI Mode?
The model appears in AI Mode’s “Ask anything” bar, accessed via the “+” icon. It sits under a “Gemini 3 models” section alongside “Auto” and “Pro.” Robby Stein, VP of Product for Google Search, confirmed the rollout in a post on X, describing the model as “better at following instructions + understanding your intent, so you get even more helpful responses.”
Access requires a Google AI Pro or Ultra subscription. The rollout is global and available in English, per Stein’s announcement. Notably, Google’s own support documentation for AI Mode’s model menu had not been updated as of Search Engine Journal’s reporting, still listing only “Fast” and “Pro” as available options, with no mention of “Auto” or the new Flash tier.
Will Gemini 3.7 Flash Become AI Mode’s Default?
Google hasn’t said. But the upgrade cycle so far has followed a consistent pattern. Gemini 3 Pro became selectable in AI Mode in November 2025. Gemini 3 Flash then became the default in December 2025. At Google I/O in May 2026, Gemini 3.5 Flash replaced it as the global default. The introduction of 3.7 Flash as a selectable, subscription-only model fits the same early stage observed in previous cycles, per Search Engine Journal’s analysis.
In February 2026, Jeff Dean, then Google’s chief scientist, stated that Flash models remain Search’s production tier because of their lower latency and cost compared to Pro models. Search Engine Journal noted that this is the cycle Google plans to continue for future model generations.
Whether 3.7 Flash feeds into “Auto” routing, the automatic model-selection system Google began building into AI Mode last November, is also unanswered. Google has not disclosed how Auto chooses between available models.
What This Means for AI-Search Visibility
Here is the simplest way to frame what’s happening: AI Mode no longer runs on a single model. Different users, on different plans, or routed to different settings by “Auto,” may now receive different answers to the exact same question. That means a brand’s presence in AI Mode responses is no longer a fixed target.
Search Engine Journal pointed to one immediate practical use: if you have noticed inconsistent citation behavior in AI Mode, you can now test whether the model is the cause. Run the same query across 3.7 Flash, Auto, and Pro, then compare which sources each one cites. That is a useful diagnostic, even if most brands currently lack an AI Pro or Ultra subscription to do it at scale.
Two implications that the sources leave largely unaddressed:
- The “Auto” routing layer is effectively a blind spot. Google has not disclosed how it selects between models. Content that earns citations from one model may not carry over to another. Brands have no reliable way to verify which model is evaluating their content when a user is on Auto.
- When official documentation lags the live product by this much, third-party tools and guides built on that documentation are likely outdated too. Any visibility data collected before August 14, 2026, and any guidance based on a fixed model assumption, should be treated with caution.
Hingewise’s read: the more significant story here is not the model upgrade itself, but the pace. Three weeks between 3.6 Flash and 3.7 Flash. A new selectable model in AI Mode within 24 hours of release. The model layer beneath AI Mode is now moving faster than most content or AI-search visibility strategies account for. Any brand monitoring its presence in AI Mode responses should treat the model version as a variable, not a constant, and build that into how they interpret their data.
What to Check Before Assuming Your AI Mode Presence Is Stable
- Confirm which model is active when you run AI Mode queries: Auto, Pro, or Flash
- Run the same branded query across all three model options and compare whether your content appears in each
- Note which citations appear across model outputs: the same page may be cited by one model and omitted by another
- Check whether your content appears as a cited source or only as a linked result below the AI response
- Test whether answers referencing your brand require follow-up prompts or appear in the first response
- Track whether Auto consistently routes to the same model across repeated queries, or switches between sessions
Google has given no timeline for when or whether Gemini 3.7 Flash will become AI Mode’s default for all users. The next inflection point to watch is whether it rolls out to the free tier, which is when its effect on AI search results at scale will become measurable.
Sources: Search Engine Journal, Google Blog, Google Search Support.
Lam Nguyen · Hingewise