Anthropic has rolled out a voice mode update extending its Opus and Sonnet models across desktop, mobile, and web apps. The update allows paid users to switch models mid-conversation and pull context from connected apps like Gmail, Slack, and Canva while supporting additional languages in beta.
Since last year, Anthropic has offered a voice mode through Claude, allowing you to speak to its chatbot instead of writing out your prompts. If I had to guess, most people probably don’t know Claude has voice input. However, for those that have used it, the consensus has been that it could use work. One issue is that before today Anthropic routed voice mode queries through Haiku, its smallest model, to reduce latency. That meant voice mode worked well enough for simple questions, but could struggle with more complicated requests. When Anthropic launched voice mode earlier this year, it was primarily focused on delivering answers to quick questions with minimal delay. But in a blog post, the company said people immediately started using voice mode for far more than casual queries. They were using it to work through real business problems,
which Haiku was not really designed for. That model kept conversations quick, but not always deep.
Model Upgrades and App Integrations
Sonnet and Opus are deeper models that the company says are designed for hard problem-solving.
They can deliver more complex responses and analysis, and then take action on your behalf. For instance, turn a conversation you have with it into a one-page pitch, or shift around your calendar appointments if your train is running late.
Users can shift back and forth between text and voice mode mid-conversation, as well as change models. Provided you pay for Claude access, the tool will default to the last system you used for text chat. You can also switch between Haiku, Sonnet or Opus mid-conversation through the model picker. Voice mode uses the fastest version of whichever model you've selected, so the conversation runs smoothly,
Anthropic notes.
Additionally, voice mode can now pull context from connected apps such as Gmail and Slack, as long as you grant Claude permission to do so, and extending its reach into apps like Canva.
Architecture and Language Expansion
An Anthropic spokesperson told Engadget voice mode uses a turn-based architecture, so all interactions will see Claude listen to you, pause to think and then respond. It’s not fully duplex like OpenAI’s new GPT-Live system, which can simultaneously process speech and generate an output. In practice, that should make talking to Claude feel less natural than ChatGPT.
Another limitation of Claude’s voice mode is that it can’t automatically detect the language you’re speaking in if you decide to switch languages mid-conversation. You need to either tell it out loud you’re about to switch or select the language you’re about to speak in from the voice settings menu. However, Anthropic has added support for additional languages, including Indonesian. Until now, languages other than English were only available in beta, but now voice mode is available in French, German, Spanish, Hindi, Indonesian, Italian, Japanese, Korean, and Portuguese.
Account Tiers and What Comes Next
Anthropic is rolling out the new voice mode in beta to all users across its desktop and mobile apps, as well as web client. If you’re using Claude through a free account, Anthropic will limit you to a single connection and your prompts will all go through Haiku, though you can speak to Claude in all of the languages voice mode now supports.
This release is focused on intelligence and tool access,
Anthropic told Engadget. We're continuing to invest in voice and we'll have more to share later this year.
Sources: theverge.com.
Worth a look
