Editorial illustration for Grab, CJ ENM, LiveKit praise Gemini 3.5 Live Translate for quality and accuracy
Grab, CJ ENM, LiveKit praise Gemini 3.5 Live Translate...
Real-time translation software has always sucked. Users gritted their teeth through a slow, robotic mess or paid out the nose for human interpreters. That changed this month.
Ride-hailing giant Grab, Korean media powerhouse CJ ENM, and video infrastructure firm LiveKit—three companies that live or die by clear communication—tested Google's new Gemini 3.5 Live Translate model. Their verdict is stark: it just works. They praise its quality, its accuracy, and a latency so low it disappears.
This isn't a tweak. It's the first time a machine has felt fast and reliable enough to use without wincing.
Today, we’re taking our next step with the release of Gemini 3.5 Live Translate, our latest audio model for live speech-to-speech translation.
Google Meet's integration details expose the old limits. It's jumping from five supported languages to over seventy. It ditches the clumsy system that forced every global chat through English as a pivot language.
Now, a single call can handle over two thousand direct language combinations. That scale changes everything. It turns the feature from an executive novelty into real infrastructure.
The praise from Grab, CJ ENM, and LiveKit matters because they've hammered it in the real world—in messy contexts with edge cases and jargon where old systems crumbled. Good translation tech should be boring. It should fade into the background.
That might finally be possible.
Related Reading
Browse all Google news →
LLMs & Generative AI ChatGPT's 'Nerdy' tweak rewards goblin metaphors in answers, study finds
LLMs & Generative AI Google tests visual 'magazine-style' UI for Gemini 3 Pro users
LLMs & Generative AI AI Engineers Face Rising Costs, Need New Strategies for Efficiency
Related Topic NVIDIA and Google Cloud let developers scale AI from prototype to production
Market Trends
Related Topic Google's FACTS benchmark shows 70% factuality ceiling across four tests
Research & Benchmarks
From Archives Apple's top AI concept mirrors vibe coding, using Shortcuts as a model
From Archives CoCoNuT paradigm expands residual stream for latent‑space, multi‑path reasoning
From Archives SpaceX inks USD 920 M/month deal with Google for 110,000 Nvidia AI chips
Business & Startups
From Archives Gemini 3.5 adds action‑taking to run complex multi‑step workflows in apps
Industry Applications
Common Questions Answered
How many languages does Gemini 3.5 Live Translate support compared to the previous Google Meet system?
Gemini 3.5 Live Translate now supports over seventy languages, a significant increase from the previous system. Additionally, it can handle over two thousand direct language combinations, eliminating the need to route all conversations through English as a pivot language.
Why is the elimination of English as a pivot language important in Gemini 3.5 Live Translate?
Removing English as a mandatory pivot language allows direct translation between any two of the supported languages, improving both speed and accuracy. This change transforms real-time translation from a clumsy workaround into practical infrastructure that can handle complex multilingual conversations seamlessly.
Which companies tested Gemini 3.5 Live Translate and why are their endorsements significant?
Grab (ride-hailing giant), CJ ENM (Korean media powerhouse), and LiveKit (video infrastructure firm) tested the new translation feature. Their praise matters because these companies depend on clear communication for their core operations and have thoroughly tested the technology in real-world scenarios.
What were the main problems with real-time translation software before Gemini 3.5 Live Translate?
Previous real-time translation software either operated slowly with robotic output quality or required users to pay for expensive human interpreters. Gemini 3.5 Live Translate addresses these limitations by providing fast, accurate translations that turn the feature into genuine infrastructure rather than an executive novelty.
