Image: Anthropic / claude.com Claude Voice Mode Now Runs on Opus and Sonnet, Can Use Your Connected Apps
Anthropic upgraded Claude's voice mode with higher-tier model support and app integrations. Voice conversations can now pull from Gmail, Slack, and Google Calendar in real time.
Anthropic rolled out a significant update to Claude’s voice mode on July 23. The feature, previously limited to the Haiku model, now supports Opus and Sonnet. More meaningfully, it can reach into connected apps like Gmail, Google Calendar, Slack, and Canva while you’re talking.
Model Choice in Voice
Until this update, voice mode ran on Haiku only. That kept responses fast but made it a poor fit for anything involving sustained reasoning or complex work. Now users can pick between Haiku, Sonnet, and Opus. The model picker is available mid-conversation, so you can start with Haiku for quick back-and-forth and switch to Opus when the discussion gets complicated.
Free accounts are restricted to Haiku. Sonnet and Opus are available on Pro and Max plans.
Connected App Access
The bigger change is tool integration. Voice mode can now call into any app you’ve connected to Claude, including Gmail, Google Docs, Google Calendar, Canva, Slack, and Notion. A spoken request like “reschedule my 3pm with the engineering team to tomorrow” can actually act on your calendar rather than just suggest what you could do.
Claude asks for explicit permission before using a connected tool during a conversation. The request comes up as a prompt and you approve or deny it before anything happens.
Free users get one connected app. Paid plans get access to all connected apps.
Languages
Voice mode now supports ten languages: English, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Portuguese (Brazilian), and Spanish. One important limitation: Claude doesn’t auto-detect your language. You need to specify it manually. Anthropic’s page notes that mixing languages in the same conversation may produce inconsistent results.
Cross-Platform Beta
The update is available in beta across mobile, desktop, and web. You start a voice conversation through the sound wave icon in the app. Usage counts toward your standard plan limits.
Anthropic’s blog post for the feature is at claude.com.
What’s Different from OpenAI
One thing Anthropic did not change: the underlying speech model. OpenAI has been iterating on voice quality, interruption handling, and conversation flow directly. This Claude update improves what the AI can do during a voice conversation (model capability, tool access), but the audio experience itself is unchanged. That’s a real distinction worth tracking as voice becomes a bigger part of how people interact with coding assistants and productivity tools.