Anthropic is making Claude voice mode more useful for work conversations, not just quick spoken replies. The company said Thursday that users can now choose between Opus, Sonnet, and Haiku models when using voice mode, expanding a feature that was previously powered by Haiku.
The update arrives weeks after OpenAI rolled out a new family of conversational models and updated ChatGPT's voice mode. Anthropic's move is different in one important way: Claude voice mode can connect to outside apps and carry out tasks, while OpenAI's updated voice mode focused on conversational style and still is not able to use different tools to get work done.
What Changed In Claude Voice Mode
Claude voice mode was released last year with Haiku as its model. That made it fast, but Anthropic said it was not well suited for complex work. The new release gives users access to Opus, Sonnet, and Haiku, which changes the role voice can play inside Claude.
Rather than treating voice as a separate, lighter experience, the updated mode follows the user's recent model choice. Anthropic said voice mode picks the last model someone used in text chat and uses its fastest version by default. That keeps the experience connected to the way the user was already working in Claude.
The change matters because spoken interaction often moves faster than typing. A user may want to rehearse an idea, test a message, or talk through a problem without turning the session into a short command. By giving voice mode access to more capable models, Anthropic is positioning it for longer and more involved exchanges.
That does not mean every user gets the same range of choices. The new voice mode is available to all users in beta across platforms, but free users are restricted to the Haiku model and only one connected app.
Why Model Choice Matters
The original Haiku-powered voice mode was built around speed. That can be useful for direct questions or brief back-and-forth exchanges, but Anthropic's own framing suggests the new update is aimed at work that needs more context and continuity.
Anthropic said the updated voice mode can help with longer conversations. The examples it gave are practical rather than casual: giving feedback on communication style, talking through a pitch to a client, and brainstorming product market research.
Those examples share a common pattern. They are not single-step requests with one short answer. They require a conversation that can develop over several turns, where the user may refine an idea, react to feedback, or ask Claude to explore alternatives.
For someone preparing a client pitch, voice can turn Claude into a real-time sounding board. For communication feedback, speaking aloud may be more natural than pasting a polished draft. For product market research brainstorming, voice can make it easier to move quickly through possibilities before deciding what is worth writing down.
App Connections Bring Voice Into Workflow
The bigger distinction in Anthropic's update is not only model selection. Claude voice mode can also tap into other apps, including Gmail, Google Calendar, Slack, Canva, and Notion.
That means a voice conversation can move beyond discussion into action. Anthropic gave examples such as updating a meeting slot, drafting an email, or creating a document in Notion. In practical terms, voice mode can become a way to manage work across connected services instead of staying confined to a chat window.
The connected-app list also shows the kinds of workflows Anthropic is targeting:
- Scheduling: Google Calendar can support requests such as changing a meeting slot.
- Communication: Gmail and Slack can support drafting or coordinating messages.
- Creation: Canva and Notion can support moving ideas into documents or visual workspaces.
This is a meaningful contrast with OpenAI's updated voice mode as described in the source article. OpenAI updated the conversational style of ChatGPT's voice mode, but it still cannot use different tools to get work done. Anthropic's update puts more emphasis on voice as an interface for task completion.
Language Support And Remaining Limits
Anthropic added multilingual support to Claude voice mode in beta earlier this year. At the moment, users can talk in various languages, but they have to manually specify the language.
The supported languages listed by Anthropic are English, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Portuguese (Brazilian), and Spanish (Latin America/Spain). That gives the beta a broader language base, but the manual language selection means the experience still depends on a user setting rather than automatic switching.
There are also limits in the voice experience itself. Anthropic did not make changes to the voice model with this release. It also has not detailed what kind of voice stack it uses.
Because of that, users should not necessarily expect the same kind of conversational improvements associated with OpenAI's release. The source article specifically notes that users might not find improvements such as better interruption handling. In other words, the main upgrade here is about model access and app-connected work, not a new voice engine.
What This Means For Claude Users
The update makes Claude voice mode more flexible, especially for people who already use Claude for writing, planning, research discussions, or workplace coordination. Instead of using voice only for quick exchanges, users can bring more complex Claude models into spoken sessions and connect those sessions to apps where work already happens.
For free users, the experience remains more limited: Haiku only, with one connected app. For everyone else covered by the beta, the key change is that voice mode can better match the model context of text chat and become part of a broader workflow.
The release also shows a clear difference in priorities between Anthropic and OpenAI's recent voice updates. OpenAI focused on conversational style. Anthropic is using this update to make Claude voice mode more capable inside day-to-day productivity tools.
For now, the feature remains in beta across platforms. The strongest reason to try it is not simply that Claude can talk back, but that a spoken Claude session can now connect model choice, longer conversations, and app actions in the same workflow.