Anthropic has updated Claude’s voice mode to support its Opus, Sonnet and Haiku models, giving users more control over the intelligence and speed behind their spoken conversations with the AI assistant.
The upgrade is designed to make Claude voice mode more useful for complex professional and creative tasks. Until now, the feature relied primarily on the lightweight Haiku model, which offered fast responses but was less suitable for work requiring deeper reasoning or extended discussion.
Claude voice mode is currently available in beta across supported platforms. The update is being introduced to both free and paid users, although access to models and connected applications will vary depending on the user’s subscription.
The updated voice experience allows users to select from Anthropic’s main Claude model families.
Opus is positioned as the most capable option for complicated reasoning, detailed analysis and demanding professional tasks. Sonnet offers a balance between performance and response speed, while Haiku is optimized for quick and efficient interactions.
Claude can also automatically use the model that a person most recently selected in text chat. Voice mode will choose the fastest available version of that model by default, helping users move between typed and spoken conversations without repeatedly adjusting their settings.
The expanded model selection could make voice mode more practical for users who want to discuss complicated ideas, review business strategies or work through detailed projects without relying entirely on text.
Anthropic says the upgraded feature can support longer and more productive conversations.
Users can ask Claude to evaluate their communication style, rehearse a presentation, prepare for a client pitch or brainstorm ideas for product and market research. The inclusion of more capable models should allow the assistant to follow detailed discussions and respond with greater depth.
This represents a shift from basic voice assistants that are mainly used for short questions, reminders or simple information requests. Anthropic is positioning Claude as a conversational work partner that can participate in extended planning and problem-solving sessions.
For professionals, voice interaction could provide a faster way to develop ideas while travelling, preparing for meetings or completing tasks that do not require direct keyboard input.
One of the most significant parts of the update is Claude’s ability to interact with connected applications.
Claude voice mode can work with services including Gmail, Google Calendar, Slack, Canva and Notion. Depending on the permissions provided by the user, the assistant can access relevant information and help complete actions through spoken instructions.
A user could ask Claude to reschedule a meeting, prepare an email, review messages from a workplace channel or create content inside Notion. This gives the voice feature a wider role than simply answering questions.
The ability to use external tools could help Anthropic differentiate Claude from competing conversational AI services. It also supports the company’s broader effort to turn Claude into an assistant that can complete practical workplace tasks across multiple platforms.
Access to connected services will still depend on account settings, subscription plans and permissions granted by users.
Claude voice mode also supports conversations in several languages.
The available options include English, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Brazilian Portuguese and Spanish for users in Latin America and Spain.
Users currently need to select their preferred language manually. Automatic language detection does not appear to be included in the latest update.
The multilingual expansion could make Claude voice mode more useful in international workplaces and regions where users regularly communicate in languages other than English.
The upgraded voice mode is available to all Claude users in beta, but the free version includes several limitations.
Free users are restricted to the Haiku model and can connect Claude to only one external application. Paid customers are expected to receive broader model access and more flexibility when using integrations.
This approach gives Anthropic a way to introduce voice features to a large audience while reserving its most advanced capabilities for subscribers.
The company has not announced major changes to the underlying voice technology in this release. As a result, users may not experience improvements in areas such as interruption handling, speech timing or conversational delivery.
The update is focused mainly on model selection, longer conversations and connected app functionality rather than changes to the way Claude processes and generates speech.
The release comes as leading AI companies compete to make voice interaction a central part of their consumer and workplace products.
OpenAI, Google, Microsoft and other technology companies have been developing conversational AI systems that allow users to communicate naturally through speech. These companies are also working to connect their assistants with productivity software, search tools and business applications.
Anthropic’s strategy appears to focus on combining advanced reasoning models with workplace integrations. By allowing Claude to use tools during voice conversations, the company is attempting to make the feature useful for completing tasks rather than limiting it to general conversation.
The updated model options could also encourage more users to rely on voice mode for complex work. Selecting Opus or Sonnet may produce more detailed responses, while Haiku remains available for faster and simpler interactions.
As AI assistants become more capable, voice interfaces are likely to play a larger role in how people manage schedules, prepare documents, communicate with colleagues and develop ideas. Anthropic’s latest Claude update is another step toward making spoken conversations a practical way to control AI-powered productivity tools.
Share your thoughts about this article.
Be the first to post a comment!
