Loading…
Google’s reported release adds visual input and extended reasoning to its live voice-agent offering, but its practical capability advantage remains unverified.
Limits of the evidence: The item is a single discovery-class source; it reports no independent evaluation or benchmark results, and says architecture and training-data details are limited.
What to watch
Assessment revised Oct 4, 2026
Voice, Video, and Reasoning in One: Google’s Gemini 3.8 Live with Extended Thinking
Andrew NgOct 2, 2026First report
First assessment, from Andrew Ng
No maturity ladder applies to this desk.
Sens.ai aggregates and assesses published reporting. The assessment above is machine generated; the original sources are authoritative.