Gemini Live Translate in Google Meet: what you get and who has it

Gemini Live Translate is Google's streaming speech-to-speech translation model, and Google Meet uses it for its speech translation feature. In Meet today, you can hear a call dubbed between English and Spanish, French, German, Portuguese or Italian, if your Workspace plan is eligible. A 70+ language version was announced in June 2026, but as of October 2026 Google describes it as a private preview for select business customers with a broader rollout "later this year". This page separates what is live from what is announced, and tells you what to use if your account is not covered.

What Gemini speech translation in Meet does today

Meet's speech translation does not subtitle the call. It generates translated audio that preserves the speaker's tone and cadence, so a German speaker is heard in English by a listener who chose English. Google made it generally available on 27 January 2026 for web, and started the mobile apps on 8 April 2026.

The launch scope is narrow, and that is the part most summaries skip:

• Languages: bidirectional translation between English and Spanish, French, German, Portuguese and Italian. Nothing between two non-English languages.
• One pair per meeting. A call with English, German and Italian speakers cannot run all three directions at once.
• Plans: Business Standard and Plus, Enterprise Standard and Plus, Frontline Plus, and the consumer Google AI Pro and Ultra tiers, plus some AI add-ons for education.
• Admin control: on by default, and switchable per organizational unit.
• Conference rooms: room hardware users can hear translations, but their own speech is not translated.

What the Gemini 3.5 upgrade changes, and what it does not yet

On 9 June 2026 Google introduced Gemini 3.5 Live Translate. The model translates continuously instead of waiting for the speaker to finish, detects the spoken language itself, and trails the speaker by a few seconds. For Meet, Google says it takes speech translation from 5 languages to 70+ and from "to and from English" to 2000+ language combinations in one meeting.

The caveat is the rollout wording: private preview for select business Workspace customers from June, broader availability later in 2026. Google has not published an announcement of general availability of the 70+ language version, so treat it as announced, not shipped, until your own admin console or the Workspace Updates blog says otherwise. Facts above are current as of 7 October 2026; check monthly.

Three products people confuse

The same model name appears in three places, and they do not do the same job.

1. Gemini in Google Meet. Translates the participants of a video call and plays dubbed audio inside the call.
2. Live translation in the Gemini app. A "Turn on live translation" setting in Gemini Live on Android, rolling out gradually server-side. It translates a face-to-face or phone-side conversation. It does not enter a Meet room.
3. Google Translate app. Gemini 3.5 Live Translate runs here too, on Android and iOS, including a listening mode that plays the translation through the phone earpiece. Again, no video call.

If you want to use a general chatbot in place of these, ChatGPT as a translator has a similar gap: it works on a phone next to you, not as a participant in a call.

Gemini in Meet vs translated captions vs a translator bot

Two notes on the table. First, Akume's own limits: it joins as a participant and the host admits it, and on Zoom the host has to allow recording for the translator to work, though no recording is kept. Second, for Meet's text option see Google Meet translation: captions, speech and bots.

Wait for the rollout or use an alternative now

Gemini in Meet is the right tool if your whole call is English plus one of the five launch languages, everyone is on eligible plans inside your organization, and you are fine with one pair per meeting. It needs no extra vendor and your admin can control it.

Waiting is reasonable if you need a pair outside that list, such as Japanese with German, and the call is not for weeks. The 70+ language version may cover it, but you cannot schedule against a "later this year" rollout.

Use an alternative now if any of these apply:

• A client or guest is outside your Workspace, and you cannot afford to find out mid-call how they are handled.
• You need three or more languages in one call. Meet speech translation runs one pair.
• You also run calls on Zoom or Teams. See Zoom live translation and Teams live translation for what those platforms offer natively.
• Your plan is not on the eligible list.

Translated captions are the zero-effort fallback, but they are text, and people listening while they work or drive still miss the call.

How to check your own account

1. Ask your Workspace admin to open the Admin console and look at the Google Meet settings for speech translation. If it is off for your organizational unit, that is a switch, not a plan limit.
2. Start a Meet on the web and look for the speech translation button in the row of call controls.
3. Run a two-person test with one internal and one external participant before any client call, and write down which direction and which pair you tried.
4. Re-check after Google's Workspace Updates posts about the 70+ language rollout.