Anthropic is upgrading Claude's voice mode, giving paid users the ability to conduct spoken conversations through its more capable Opus and Sonnet models for the first time.

Until this release, every voice request in Claude was routed through Haiku 4.5, Anthropic's smaller and faster model. The change is designed to make voice mode better suited for more complex tasks and comes only weeks after OpenAI released its own conversational platform upgrade, GPT-Live. Both companies are competing to establish voice as a primary interface for AI assistants, a space where Apple's Siri, Amazon's Alexa, and Google Assistant have also been integrating large language model capabilities.

Claude voice mode moves beyond Haiku

Voice mode has been available in Claude since last year, allowing users to speak to the chatbot instead of typing. As Engadget noted, Anthropic had been routing spoken requests through Haiku, its smallest and fastest model, in order to keep response times low. That approach worked for quick questions but was less effective for more demanding prompts.

The new update lets users choose Opus or Sonnet for voice conversations. Anthropic told TechCrunch that voice mode now defaults to whichever model a user most recently used in text chat, while running the fastest version available so the exchange remains fluid. According to 9to5Mac, a model picker inside the interface allows users to switch among Haiku, Sonnet, and Opus without leaving the conversation.

TestingCatalog, which detected the feature under a flag before the official announcement, reported that the picker had appeared inside voice mode for about three weeks as a visual control only. During that period, every session was still routed to Claude Haiku 4.5 regardless of which model was selected.

The selection of Opus or Sonnet began changing the actual model only this week.

App connectors bring voice mode into workplace tasks

The upgrade also gives voice mode access to Claude's app connectors. TechCrunch reported that the assistant can now interact with Gmail, Google Calendar, Slack, Canva, and Notion, enabling spoken requests to reschedule a meeting, draft an email, or create a Notion document. Engadget also confirmed that voice mode can draw context from connected apps after a user grants permission.

Anthropic described the update as enabling longer and more complicated tasks, such as rehearsing a client pitch, receiving feedback on communication style, or working through product research aloud. "This release is focused on intelligence and tool access," the company said, adding that it plans further voice-related work later this year.

As Cryptopolitan previously reported, Anthropic reached a valuation of about $350 billion in November 2025 and generates approximately 80% of its revenue from more than 300,000 business and enterprise customers.

That enterprise base makes connector access a significant part of the voice upgrade. Business users can ask Claude questions about emails, calendars, Slack messages, and documents using voice queries, and the assistant can take actions within those connected services rather than only providing spoken answers. The feature also aligns with how larger customers have deployed Claude, including autonomous agents used across 27 ministries in Alberta and by Deloitte, which has 470,000 employees.

Anthropic broadens voice language support

9to5Mac listed 11 supported languages for Claude voice mode: English, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Brazilian Portuguese, and Spanish for both Latin America and Spain.

Switching between languages does not happen automatically. Engadget reported that users must either say aloud that they are changing languages or choose a different language in the voice settings menu. Free accounts remain limited to Haiku and one connected app, but they can still use all supported languages.

Claude and GPT-Live use different voice approaches

Anthropic and OpenAI are taking different approaches to voice interaction. According to an Anthropic representative, Claude uses a turn-based system, meaning it listens, thinks, and then speaks. OpenAI's GPT-Live, by contrast, can listen and speak at the same time. GPT-Live was released to all ChatGPT subscribers on July 8, 2026, and was described as "the primary interface to computing."

TechCrunch reported that Anthropic did not change Claude's underlying voice model in this release and has not fully explained the technology behind its voice stack. As a result, users may not notice major improvements in areas such as smoother interruptions.

TestingCatalog reported that Claude still appears to rely on text-to-speech technology for voice output, with speech generated through ElevenLabs. Even so, TestingCatalog said interruption handling continued to work properly.

Tool access remains one area where Anthropic's voice mode differs from OpenAI's refreshed voice experience. TechCrunch noted that OpenAI's updated voice mode changed its conversational style but still cannot use tools to complete tasks. That distinction means Claude can act on a user's behalf within third-party apps by voice, while OpenAI's voice mode remains limited to conversation. Anthropic has said it plans further voice-related updates later this year, though it has not detailed what those will include.

The updated Claude voice mode is rolling out in beta to all users across desktop, mobile, and web.