Claude Voice Mode Now Supports Opus and Sonnet: What Changed for Professionals (July 2026)

TL;DR. Claude's voice mode, previously limited to the Haiku model, now lets paid users switch to Sonnet or Opus mid-conversation. It also gained app integrations (Gmail, Slack, Calendar, Canva, Notion) and 10-language support in beta. Free users stay on Haiku with one connected app.

---

Since voice mode launched for Claude, it had one limitation that frustrated anyone trying to do real work: it only ran on Haiku. Haiku is fast — designed for instant responses — but it's not the model most paid Claude users choose for anything that requires sustained reasoning. If you were mid-document, mid-analysis, or mid-research session in Sonnet or Opus and switched to voice, you dropped to a less capable model with no warning.

On July 23, 2026, Anthropic changed that.

What's new

Model choice. Voice mode now supports Haiku, Sonnet, and Opus. You can switch between them during a conversation using the model picker, or just leave it on whatever you were last using in text chat — Claude picks up automatically from your prior session's model selection. Anthropic says it uses "the fastest version" of whichever model you choose, which means some speed optimization is applied even for Opus, though it will still feel slower than Haiku in quick back-and-forth exchanges.

App integrations. Voice mode can now pull context from and take action in connected apps: Gmail, Google Calendar, Google Docs, Slack, Canva, and Notion. In practice this means you can say "find the email from Sarah about the contract and summarize it" or "check my calendar Thursday afternoon and draft a decline for anything after 3pm." The apps are the same connectors already available in Claude's text interface — voice just gains access to them.

Language support. Ten languages are now in beta: English, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Portuguese (Brazilian), and Spanish (Latin America and Spain). You need to specify your language manually — Claude won't detect it automatically — but mid-conversation switching is supported.

What didn't change

It's worth being precise about what this upgrade is and isn't. Anthropic made no changes to the voice model itself — the speech layer that handles audio in and out. You won't see improvements in interruption handling, conversational naturalness, or speaking pace. Those remain the same as before. What changed is the intelligence running underneath: the model that reads your words and generates a response. Haiku was doing that job before; Sonnet and Opus can do it now.

This is a meaningful but specific improvement. If you were frustrated that voice mode felt shallow — that it couldn't reason through a complex question or maintain a long thread — that's now fixed on paid plans. If you were frustrated that voice mode sometimes talked over you or felt robotic in cadence, that's still the same.

Free vs. paid split

	Free plan	Paid plans (Pro, Max, Team, Enterprise)
Available models	Haiku only	Haiku, Sonnet, Opus
Connected apps	One app	Multiple apps simultaneously
Claude Fable 5	Not available	Not available (no plan)
Languages (beta)	All 10	All 10

Claude Fable 5 — currently the most capable generally available Claude model — is not supported in voice mode across any plan. Anthropic has not announced a timeline for adding it.

What this looks like in practice

The upgrade matters most for professionals whose work involves both extended reasoning and hands-free interaction. A few realistic scenarios:

Research by voice. You're reviewing background for a client or case and want to think out loud. On Haiku, Claude could track what you said but struggled with multi-step synthesis. On Sonnet or Opus, you can walk through a complex question verbally and get an answer at the level you'd expect from text — useful in a car, on a walk, or between meetings.

Voice-driven scheduling and comms. With Calendar and Gmail connected, you can handle routine coordination entirely by voice: "Do I have anything before noon tomorrow? Block two hours in the afternoon for focused work. And draft a follow-up to the thread with Marcus from last week." This is the most immediately practical use for most professionals — not deep reasoning, but efficient task dispatch without opening a laptop.

Multilingual work. For non-English speakers or professionals working across language contexts, the beta languages make it possible to conduct a voice session in French, Japanese, or Spanish. Quality varies by language (English still has the best support), but it works.

Things to know before you switch

• Voice conversations count against your standard usage limits, same as text.
• Text transcripts are saved automatically after a voice session.
• Using multiple connected apps simultaneously may cause delays — Claude has to make multiple tool calls in sequence.
• Voice mode is not available in Claude Cowork or Claude Code.
• Enterprise admins can disable voice mode entirely for their organization.

The takeaway

Claude's voice mode was useful but constrained before this update — useful for quick questions, constrained for anything requiring Sonnet or Opus-level reasoning. The July 23 upgrade removes that ceiling for paid users. Whether voice is the right interface for your work is still a personal question, but the model limitation is no longer the reason to stay away.

If you're a Pro, Max, or Team subscriber who has avoided voice mode because it felt like a step down from your usual text experience, it's worth trying again.

---

Sources

• Anthropic updates Claude voice mode with more capable models — TechCrunch
• Claude's voice mode is now available for Opus and Sonnet — The Verge
• Anthropic upgrades Claude voice mode with more powerful models — 9to5Mac
• Claude's voice mode just got smarter — Engadget
• Use voice mode — Anthropic Help Center