Artificial Intelligence

Anthropic upgrades Claude voice mode to Opus, Sonnet, and Haiku — and gives it app control

Anthropic upgraded Claude's voice mode to Opus, Sonnet, and Haiku and gave it control of Gmail, Calendar, Slack, and more — a step beyond OpenAI's voice.

Anthropic is making Claude’s voice mode substantially more capable, announcing Thursday that users can now choose between its Opus, Sonnet, and Haiku models when talking to the assistant, according to TechCrunch. The move comes weeks after OpenAI rolled out a new family of conversational models for ChatGPT’s voice experience.

Until now, Claude’s voice mode ran solely on Haiku, Anthropic’s fastest model — good for quick responses but not suited to complex work. In a blog post cited by The Verge, the company said people immediately began using voice for far more than casual queries, working “through real business problems” that Haiku “kept conversations quick, but not always deep.” With the update, voice mode defaults to the fastest version of whichever model a user last used in text chat.

The strategic differentiation is agency. Claude’s voice mode can now tap into connected apps like Gmail, Google Calendar, Slack, Canva, and Notion — so users can ask it to move a meeting, draft an email, or create a document by speaking, per TechCrunch. As the outlet noted, that is a big difference from OpenAI’s voice mode, which updated its conversational style but still can’t use tools to get work done.

Claude Voice Mode Becomes an Agent You Talk To

Anthropic says the upgraded voice mode suits longer conversations: rehearsing a pitch to a client, getting feedback on communication style, or brainstorming product market research, per TechCrunch. Sonnet and Opus are “designed for hard problem-solving,” the company said in the post quoted by The Verge — able to deliver deeper analysis and then act on it, like turning a spoken conversation into a one-page pitch or shifting calendar appointments when a train runs late.

Users can switch between text and voice mid-conversation and change models on the fly — a quick Haiku chat can escalate seamlessly into an Opus deep-dive, The Verge reported.

Multilingual support is also expanding beyond the earlier beta. Anthropic supports English, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Portuguese, and Spanish, though users must manually specify their language, per TechCrunch.

What Anthropic Didn’t Change

Notably, Anthropic made no changes to the underlying voice model itself and has not detailed what voice stack it uses, TechCrunch reported. That means users shouldn’t expect the conversational polish OpenAI shipped — such as better interruption handling — since the upgrade is about intelligence and tool use, not speech quality.

The rollout is broad but tiered: the new voice mode is available to all users in beta across platforms, but free users are restricted to the Haiku model and a single connected app, per TechCrunch.

The update sharpens the contrast between the two labs’ assistant strategies at a moment when Anthropic is riding enterprise momentum — including AMD’s multi-billion dollar bet on its compute demand — while its models increasingly anchor the broader scramble for AI capacity. Voice, in Anthropic’s framing, is just another interface to an agent that does work — not a separate product.

FAQ

Which models power Claude’s voice mode now?

Users can choose between Opus, Sonnet, and Haiku; voice mode defaults to the fastest version of the model they last used in text chat.

What apps can Claude voice mode control?

Connected apps include Gmail, Google Calendar, Slack, Canva, and Notion — users can draft emails, move meetings, or create documents by voice.

Is the upgraded voice mode free?

It is rolling out to all users in beta, but free-tier users are limited to the Haiku model and only one connected app.