Key takeaways

  • Google announced 66 new regional voices on November 16, 2026, rolling out in the Gemini app on Android and iOS.
  • Named options include U.S. English, U.K. English, India English, Hindi, Brazilian Portuguese, Indonesian and Korean. That list is not a claim of 66 languages.
  • Google says the selected voice applies across Gemini Live and chat’s Listen feature; users can preview voices in Live’s three-dot menu or app settings.
  • The announcement supplies no new API model ID, pricing schedule or quality benchmark. We did not listen to or test the voices.
01

What Google announced

Google announced on November 16, 2026 that it is rolling out 66 new regional voices in the Gemini app on Android and iOS. The company names U.S. English, U.K. English, India English, Hindi, Brazilian Portuguese, Indonesian and Korean among the supported options, with tailored choices for each. This is a mobile AI product update, not an announced developer model or API launch.

The announcement comes from Neel Joshi, Google’s director of product management for Gemini Live. Its publication metadata gives November 16 at 18:45 UTC. Google says the update follows plans shared at Google I/O to bring local dialects to Gemini; the news now is the rollout, rather than that earlier promise.

02

One voice selection across the app

According to Google, a selected voice carries across the entire app: talking in Gemini Live, practicing a language, and tapping the Listen icon in chat to hear a response read aloud. The company describes the voices as tuned for natural cadence and everyday clarity rather than exaggerated accents. Those are vendor descriptions, not listening-test findings from AccessAllGPT.

To preview the options, Google directs users to the three-dot menu in Gemini Live or to Gemini’s voice in app settings. Readers can check there on their intended Android or iOS account. Because this is a rollout, the announcement does not establish that every account already has all options.

03

What the announcement does not establish

Sixty-six voices does not mean 66 languages, and a regional speaking voice is not evidence of improved speech recognition, translation or factual accuracy. The post lists examples rather than a complete voice-by-language inventory. It does not provide an accent-comprehension study, pronunciation benchmark or latency measurement.

The reviewed announcement does not specify a required subscription, separate charge, usage quota, age threshold or country-by-country eligibility matrix. We therefore do not describe the new voices as free, unlimited or available to every user. It also gives no API identifier or endpoint for developers to select these voices in their own products.

Do not infer voice cloning, custom voice creation or downloadable speech models from this release. The documented change is a selection of regional voices inside the Gemini app. An app voice option is not a deployment specification for a contact center or an API-based assistant.

04

Why this matters, and what to try next

Our analysis: a familiar speaking style can make a conversational assistant more comfortable to use, especially when a team has treated a single default voice as suitable for everyone. The practical change is choice across spoken interaction and read-aloud responses, without having to treat them as separate workflows. Whether that choice improves intelligibility for a particular listener still needs a real trial.

If the options are present, compare two voices using the same short, non-sensitive response. Include local place names, dates, amounts and one sentence you commonly need to hear aloud. Then check Gemini Live and chat’s Listen feature separately. Listen for understandable pronunciation and pace rather than assuming a regional label guarantees either. This is a suggested comparison, not a test we performed.

For an organization choosing a voice product, keep this app update separate from API procurement. Wait for developer documentation if the requirement is a stable voice identifier, integration pricing, service limits or contractual availability. For an individual Gemini user, the immediate next step is simpler: preview the regional choices in settings and keep the one that is easiest to understand.

05

A short voice-selection check

Suggested checks for a personal, non-sensitive trial; no results are claimed.

Entries stay in this browser tab and are not submitted to AccessAllGPT. Blank responses are copied as [Unresolved].

Check the actual choices in the intended Android or iOS account rather than assuming the rollout is complete.

Compare the same place names, dates and amounts with two voices; note pronunciation and pace.

Check both Gemini Live and chat’s Listen feature after changing the voice.

Do not substitute app availability for a documented API voice, price or service commitment.

Primary sources

  1. Regional voices are rolling out in the Gemini appGoogle · Reviewed: Full November 16 announcement and NewsArticle metadata; 66 regional voices; named language and regional examples; app-wide selection behavior; Android and iOS rollout; Live menu and settings instructions · Retrieved · Supports: Primary evidence for the announced voice count, example regional options, supported mobile platforms and selection workflow. Naturalness is a Google claim; this source does not establish independent quality, API access, pricing or complete account eligibility.

Limitations

One first-party announcement, retrieved directly. No authenticated app access, audio playback, screenshots of the product, listening study, performance benchmark or billing test. Named regional options are examples, not a complete inventory. Rollout completion, exact account eligibility, pricing, quotas and API availability were not established. The proposed comparison is editorial guidance, not observed product behavior.

Disclosures

AccessAllGPT received no payment, private access or vendor review for this briefing. Google did not review or endorse it. AccessAllGPT Research is operated by NeuralArc and is independent of Google. Publication-wide relationships are listed on the disclosures page.

Further AccessAllGPT guidance

  1. Automation workflows and human approval boundaries
  2. Gemini Notebook adds interactive sites
  3. AI automation
  4. How this publication is made