Opera 公司推出了名为 WebAI 的新网络标准，允许用户直接在浏览器中本地运行 AI 模型，无需依赖云端 1。该标准已在 Opera One 浏览器（旗舰版）中作为早期预览版发布，后续将覆盖 Opera GX 和 Opera Neon 1。
WebAI 支持包括 Google Gemma（谷歌两年前推出的开放权重模型）和 Apple 端侧基础模型在内的多种开源权重 AI 系统 1。通过这一标准，用户可选择加入（opt-in），数据保留在本地设备，无需订阅云端 AI 服务，开发者也无需维护 AI 基础设施 1。然而，本地运行会增加计算机负担，性能取决于硬件配置 1。
Opera has introduced a new web standard named WebAI, which allows users to run artificial intelligence models directly within their browsers without relying on cloud infrastructure 1. The standard is currently available as an early preview in the flagship Opera One browser, with plans to extend support to Opera GX and Opera Neon 1. By facilitating local processing, WebAI aims to enhance user privacy controls while removing the necessity for subscriptions to cloud-based AI services 1.
The framework supports various open-weight AI systems, including Google’s Gemma model—which was launched two years ago—and Apple’s on-device foundation model 1. This architecture benefits developers by eliminating the need to maintain separate AI infrastructure, while offering users an opt-in mechanism that ensures data remains stored locally on their devices 1. However, local execution imposes additional computational burdens on hardware, meaning performance ultimately depends on the specific configuration of the user's device 1.
评论
还没有评论，欢迎留下第一条。