Anthropic upgrades Claude voice mode with model choice and app integrations

Until this week, Claude’s voice mode ran exclusively on Haiku, the fastest and lightest model in Anthropic‘s lineup. The company changed that on Thursday, making Opus and Sonnet available to paid subscribers. Voice mode now also inherits context from the current text session — if you were last working with Sonnet in chat, voice mode opens with Sonnet already active. The feature released in beta across iOS, Android, and web.

The difference between the three models is real enough to be worth selecting deliberately. Opus handles the heaviest reasoning tasks — extended analysis, nuanced brainstorming, communication coaching before a client pitch. Sonnet covers most everyday conversations with better accuracy than Haiku at manageable latency. Haiku remains the default for free users and anyone prioritizing speed. The same model logic that applies to text now applies to voice.

The more consequential addition is app connectivity. Paid subscribers can link Claude’s voice mode to Gmail, Google Calendar, Slack, Canva, and Notion. A voice command can update a calendar event, draft a reply in Gmail, or create a document in Notion without switching context. Anthropic is pitching this as hands-free productivity — the kind of workflow that makes sense between meetings or away from a keyboard.

The update arrived weeks after OpenAI changed its own voice mode. The key distinction is what each can do: OpenAI’s voice mode converses, but Claude’s can take actions inside connected apps. Anthropic is positioning that operational reach as the differentiator. No improvements to the underlying voice generation model were announced — the quality of the audio output itself is unchanged from the previous version.

Free-tier users face a meaningful gap. The free plan restricts voice mode to Haiku and limits app connections to one service rather than the full five. Multilingual support covers ten languages — English, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Portuguese (Brazilian), and Spanish — but the system requires users to specify the language manually. Automatic language detection is not included.

Claude voice mode is available now on iOS, Android, and web. The model selector and full app connectivity are in beta for paid subscribers. Anthropic has not announced a timeline for general availability or what conditions would move the features out of beta.

Tags: Anthropic, Claude, voice AI, Claude voice mode, Haiku, Opus