On July 23, Anthropic announced an overhaul of the voice mode for its conversational assistant, Claude. Until now, voice mode was locked to the fast-responding Haiku model, but the update lets users pick from three models—Opus, Sonnet, and Haiku—depending on the task. It also connects with major apps such as Gmail, Google Calendar, Slack, Canva, and Notion, so you can reschedule a meeting or draft an email entirely by voice. The new voice mode is rolling out in beta across platforms.

Faster and Smarter: A Voice Mode That Lets You Choose the Model

Claude's voice mode, which launched last year, used the lightweight Haiku model to return quick responses, but it fell short on complex work. The update removes that limitation by letting you choose among Opus, Sonnet, and Haiku. By default, it inherits the model you were last using in text chat and automatically applies that model's fastest version. Because you can switch models mid-conversation, you can use Haiku for light exchanges and Opus for questions you want to think through carefully.

Anthropic says that being able to choose a more capable model makes voice mode better suited to longer conversations. Examples it highlights include getting feedback on how you speak and communicate, rehearsing a client pitch out loud, and brainstorming ideas for a new product or angles for market research. As a partner for talking through half-formed thoughts, voice mode has become noticeably more practical.

Connecting to Major Apps to Get Work Done by Voice

Another highlight is integration with outside apps. Claude's voice mode can now reach services like Gmail, Google Calendar, Slack, Canva, and Notion. That means you can adjust a meeting time, have it draft an email, or set up a new document in Notion—all by voice. Being able to operate everyday work apps by voice, rather than just answering questions, ties directly into day-to-day usefulness.

This also stands out as a difference from rival OpenAI's voice mode. OpenAI recently revamped the conversational style of its voice mode, but it still cannot operate external tools to actually complete tasks. Anthropic has clearly staked out a direction: a voice assistant you can use for real work, moving tasks forward as you talk.

Support for 10 Languages Including Japanese; Limits on the Free Plan

Language support is expanding, too. The new voice mode supports 10 languages: English, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Portuguese (Brazil), and Spanish (Latin America/Spain). However, users must specify the spoken language manually. Anthropic added multilingual support to voice mode in beta earlier this year, and this release carries that forward.

Availability is in beta across platforms and open to all users. That said, the free plan is limited to the Haiku model and a single connected app. Making full use of Opus, Sonnet, and multiple app integrations requires a paid plan.

The Voice Itself Is Unchanged—A Contrast with OpenAI on Conversation Quality

On the other hand, this update focused on model selection and app integration; the voice generation itself was not changed. Anthropic has not detailed what kind of voice stack it uses, either. As a result, improvements to the naturalness of conversation—such as better handling of interruptions mid-conversation—are unlikely to arrive with this update. That is a contrast with OpenAI's recent update, which put its emphasis on revamping conversational style. Even so, the wider choice of models and the app integration—both practical, work-oriented gains—look like a clear step forward for users who want to fold a voice assistant into their daily tasks.

Summary

Anthropic has overhauled Claude's voice mode, letting users choose among three models: Opus, Sonnet, and Haiku. It inherits the model you last used in text chat and can be switched mid-conversation. A key feature is integration with Gmail, Google Calendar, Slack, Canva, and Notion, letting you reschedule, draft emails, and create documents by voice. It supports 10 languages including Japanese and is offered in beta across platforms, though the free plan is limited to Haiku and one connected app. The voice itself is unchanged, making this an update focused on practical usefulness rather than the naturalness of conversation.