Anthropic Expands Claude Voice Mode With Multiple AI Models and App Integrations

Anthropic unveiled a significant enhancement to Claude's voice functionality, expanding what the feature can do and which AI models power it. The update arrives as rivals like OpenAI continue strengthening their own voice-enabled conversational tools.

New model options and smarter defaults#

Previously, Claude's voice mode relied exclusively on the lightweight Haiku model, which could handle simple queries quickly but struggled with nuanced or extended interactions. The fresh rollout lets users select from three tiers: Opus (most capable), Sonnet (mid-range), and Haiku (fastest). By default, the system will use whichever model someone last engaged with in text mode, switching to its quickest version to maintain responsiveness.

This flexibility means users tackling deeper work—refining a client pitch, unpacking research findings, or getting real-time feedback on how they're communicating—now have access to more powerful reasoning without sacrificing speed.

Bridging voice and productivity tools#

A notable differentiator from OpenAI's voice offering is Claude's ability to interact with external applications. Users can now connect voice mode to Gmail, Google Calendar, Slack, Canva, and Notion, enabling tasks like rescheduling meetings, drafting emails, or spinning up documents through natural conversation rather than switching between tools.

Multilingual expansion and availability#

Anthropic rolled out multilingual support for Claude's voice mode earlier this year and has continued broadening language coverage. The system currently handles English, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Portuguese (Brazilian), and Spanish (Latin America and Spain)—though users must manually select their language.

The new voice mode is now available to all users in beta across platforms. Free-tier users face restrictions: they can only access the Haiku model and connect one app integration.

What's not changing#

Anthropic made no alterations to the underlying voice model itself in this release and hasn't revealed specifics about its voice technology stack. That means users may not see the conversational refinements OpenAI recently introduced—such as improved interrupt detection or more natural dialogue flow.