Audio interactions with artificial intelligence often suffer when lightweight models fail to handle intricate questions or detailed context. Anthropic is addressing this capability gap by rolling out an update that allows Claude’s voice feature to operate on high-performing AI models alongside its base engine.
This update enables Claude Voice Mode to run on Sonnet and Opus, expanding beyond the lightweight Haiku model. Available to paid subscribers, the upgrade allows spoken prompts to execute complex reasoning, control connected apps like Gmail and Slack, and toggle between multiple global languages within a continuous conversational thread.
Upgrading Beyond Haiku to Advanced AI Models
Previously, Claude’s spoken voice interactions relied exclusively on Haiku, the smallest and fastest model in Anthropic’s lineup. While Haiku minimized latency for realistic verbal pacing, its reduced parameter size made it prone to errors when handling detailed problem-solving or multi-step logic.
With the latest system patch, users can select higher-tier models—specifically Sonnet and Opus—before launching voice mode via the waveform icon. Accessing Opus brings deep analytical capabilities to spoken queries, making it easier to discuss technical topics, analyze strategy, or work through complex problems hands-free.
Seamless Workflow Continuity and App Integration
The updated voice interface maintains context across communication channels. Users can begin a project by typing a prompt, switch to voice mode mid-session, and resume typing without losing conversation history or needing to re-explain background details.
Voice commands can also trigger actions within integrated software tools, including Google Docs, Gmail, Google Calendar, Slack, Notion, and Canva. For safety, Claude prompts the user for verbal confirmation before finalizing actions like sending emails or updating scheduled meetings.
Expanded Language Support and Plan Availability
Anthropic has also broadened Claude’s multilingual capacity. Spoken interactions now support French, German, Hindi, Indonesian, Italian, Japanese, Korean, Brazilian Portuguese, and Spanish (both European and Latin American regional variants). Users can instruct Claude to switch languages dynamically during an active conversation.
Access to these advanced features depends on account subscription tiers. While free users receive full language support, model selection remains restricted to Haiku, and app connectivity is capped at a single integrated tool.
Claude Voice Mode Capability Comparison
| Feature Tier | Free Plan | Paid Plans (Pro / Team / Enterprise) |
|---|---|---|
| Supported Models | Haiku only | Haiku, Sonnet, and Opus |
| App Integrations | 1 Connected App | Full Workspace Access (Gmail, Slack, Notion, Canva, etc.) |
| Language Support | All 10+ Supported Languages | All 10+ Supported Languages |
| Workflow Continuity | Standard Text/Voice Switching | Standard Text/Voice Switching (Counts toward normal usage limits) |
Frequently Asked Questions
How do I switch AI models in Claude Voice Mode?
Select your desired model from the top dropdown menu on the chat screen before opening voice mode. Once chosen, tapping the waveform icon initiates spoken interaction using that specific engine.
Which languages are currently supported in Claude Voice Mode?
Claude supports spoken conversations in French, German, Hindi, Indonesian, Italian, Japanese, Korean, Brazilian Portuguese, and Spanish. Language preferences can be set in Settings or changed mid-conversation via spoken commands.
Can free tier accounts use Opus or Sonnet with Voice Mode?
No, free accounts are restricted to the Haiku model and limited app connectivity when using voice features. Upgrading to a paid plan is required to unlock Sonnet and Opus for voice interactions.