Anthropic brings Opus 5 and Sonnet 5 to Claude's voice mode

11:15 AM · July 28, 2026

Anthropic updated Claude's voice mode on July 23, 2026 to support its more capable Opus and Sonnet models in addition to the Haiku model the feature launched with, giving paying subscribers the option to trade speed for depth during a spoken conversation. The system defaults to whichever model a user last used in text chat, and free users remain limited to Haiku with a single connected app. Beyond model choice, the update lets voice mode tap into a user's connected services, so someone can ask Claude by voice to update a meeting slot, draft an email, or create a document in Notion, and have the action actually carried out rather than just described. Anthropic pitched the changes around longer, more substantive conversations, such as getting feedback on communication style, talking through a pitch to a client, or brainstorming product market research, rather than the short transactional exchanges voice assistants have traditionally handled. The rollout supports ten languages and remains in beta across mobile, desktop and web, though it does not extend to Claude Cowork or Claude Code. Notably, Anthropic did not add real time model level voice improvements alongside the update, a contrast with OpenAI's concurrent voice mode work, leaving Claude's spoken interaction still built on text to speech and speech to text layered over its chat models rather than a native audio model.