Anthropic Expands Claude’s Voice Mode to Opus and Sonnet Models

Anthropic has significantly upgraded its conversational capabilities by bringing Claude’s voice mode to its most advanced reasoning models. This expansion transitions the feature from a basic utility to a high-intelligence assistant capable of handling complex tasks across mobile, desktop, and web platforms.

Powering Voice with Opus and Sonnet Intelligence

Previously, Claude’s voice mode was limited to the lightweight Haiku model, which prioritized speed over deep reasoning. In a major architectural shift, Anthropic now allows users to run voice interactions on its flagship Opus and Sonnet models. This means that instead of simple command-and-control interactions, users can engage in high-level brainstorming, complex problem-solving, and nuanced reasoning using only their voice.

Crucially, Anthropic has introduced model flexibility, allowing users to switch between different models mid-conversation. This enables a seamless workflow where a user might start a task with a fast model and then switch to the more capable Opus to refine a sophisticated concept, all without leaving the voice interface.

Multilingual Support and Tool Integration

The updated voice mode is not just a localized improvement but a global one, supporting eleven different languages. However, the true differentiator for Anthropic lies in its ecosystem connectivity. Unlike many competitors that focus solely on conversational fluidity, Claude is leaning heavily into "agentic" behavior through tool integration.

Users can grant Claude access to external productivity tools such as Gmail, Google Calendar, and Slack. This allows for a highly functional hands-free experience: a user can verbally dictate a complex project update, ask Claude to summarize it, and then command the AI to compose and send an email via Gmail. Anthropic currently holds a competitive edge in this specific niche, being the only provider that enables users to compose and save emails directly from the audio interface.

The Competitive Landscape: Claude vs. OpenAI and Google

The evolution of voice AI is currently a three-way race between Anthropic, OpenAI, and Google, each taking a different technical approach. OpenAI’s GPT-Live utilizes full-duplex audio, allowing it to listen and speak simultaneously for a more human-like flow. Google’s Gemini Live follows a similar philosophy but remains largely confined to smartphone environments.

Claude, by contrast, utilizes a turn-based system where the model waits for the user to finish speaking before responding. While this may feel slightly less "fluid" than the full-duplex models, Anthropic is betting that utility will triumph over pure conversational mimicry. By focusing on the integration of LLM reasoning with real-world software tools, Claude is moving away from being just a chatbot and toward becoming a voice-activated productivity agent.

Key Takeaways

  • Model Upgrades: Claude’s voice mode is no longer restricted to Haiku; it now supports the highly capable Opus and Sonnet models across web, desktop, and mobile.
  • Actionable Intelligence: Anthropic distinguishes itself through deep tool integration, allowing users to manage Gmail, Google Calendar, and Slack via voice commands.
  • Strategic Pivot: While OpenAI leads in conversational "naturalness" through full-duplex audio, Anthropic is focusing on the "agentic" utility of voice-driven workflows.