// AI news · 2026-10-27 · Google AI Changelog
Google introduces swappable inference formats for A2UI
Google introduces swappable inference formats for A2UI, with A2UI Express reducing latency by up to 3x and token usage by up to 3x.
Published 2026-10-27 08:00 (Beijing time)Source Google AI ChangelogAuto-summarised, fact-checked against source
01Summary
The format uses a Python-like syntax and is compiled into standard A2UI messages via SDK. Jev enables fast layout selection for predefined screen structures, improving reliability and speed. Available in Python, TypeScript, and Dart SDKs, users can choose inference formats based on use case.
Summarised from the original article by an open-source model, then checked automatically: every number and English name must appear in the source. The original article is authoritative.
02Key points
- A2UI Express reduces latency by up to 3x and token usage by up to 3x.
- Jev enables fast layout selection for predefined screen structures.
- Inference format is derived from schema and compiled via SDK into valid A2UI messages.
- Available in Python, TypeScript, and Dart agent SDKs.
03Source
Fast, reliable A2UI, with swappable inference formats
Google AI Changelog · published 2026-10-27 08:00 Beijing time
05Related news
- Conversational Commerce Prototype with Multi-Agent Swarms on GCP2026-10-27 · Google AI Changelog
- Apple Releases iOS 27.2 with Apple Pencil and Edge Light2026-10-27 · cnBeta
- Gemini Robotics ER 2 Now Available on Google Cloud2026-10-27 · Google Cloud AI/ML
- Eight Examples of Gemini 4 Argon in Action2026-10-27 · Google Gemini
- Google proposes 'sleep framework' for LLM continual learning2026-10-28 · Google Research Blog
- Oracle AI Database@Google Cloud Adds Gemini Integration and Agentic Capabilities2026-10-28 · Google Cloud
From AIMap AI news: collected from 1,100+ official and media sources. Summaries only restate the original article; follow the link for the full story. All AI news · Today