Anthropic upgrades Claude voice mode with Opus, Sonnet, Haiku models and app integrations

Anthropic is making its move in the voice AI space, weeks after OpenAI rolled out new conversational models and updated ChatGPT's voice capabilities. The company announced Thursday that Claude's voice mode now supports three model tiers — Opus, Sonnet, and Haiku — rather than being limited to the smaller Haiku model.

Previously, Claude's voice mode was powered exclusively by the Haiku model, which offered fast responses but struggled with complex tasks. The new update lets users choose between models, and Anthropic says the voice mode automatically picks the last model used in text chat, selecting its fastest variant.

The most significant upgrade is the integration with external applications. Claude's voice mode can now connect to Gmail, Google Calendar, Slack, Canva, and Notion, enabling users to reschedule meetings, draft emails, or create documents through voice commands alone. This gives Claude a notable edge over OpenAI's voice mode, which upgraded its conversational style but still lacks the ability to use external tools to get work done.

Anthropic says the new voice mode handles longer conversations more effectively, including providing feedback on communication style, talking through client pitch rehearsals, and brainstorming product market research. These capabilities push voice AI beyond simple Q&A toward genuine productivity assistance.

Multilingual support was added in beta earlier this year, and now supports English, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Portuguese (Brazilian), and Spanish (Latin America/Spain). Users must manually specify the language they wish to use.

The new voice mode is available in beta across all platforms to all users. Free-tier users are restricted to the Haiku model and can connect only one external app.

Notably, Anthropic did not update the underlying voice model itself or disclose details about its voice technology stack. This means users may not see improvements in conversational handling such as better interruption management, which OpenAI's latest voice update delivered.