1 hour ago
19
Anthropic has been quietly shipping improvements to Claude’s Voice Mode, and the latest round of updates makes the experience feel meaningfully more polished. The refresh includes a model selector inside the voice interface, a glowing animation around the central voice orb during active sessions, and a push-to-talk option that gives users more control over when the microphone is live.
What actually changed
The most practically useful addition is the model selector. Previously, Voice Mode defaulted to a fixed model, with Claude Haiku 4.5 as the prominent option in the interface. Now users can choose between different Claude models within the voice session itself.
The glowing orb animation is a smaller but notable UX touch. During an active voice session, the central orb now pulses with a visual glow to signal that the model is speaking or listening.
Push-to-talk is the third piece. Rather than relying entirely on automatic voice activity detection, users can now hold a button to control when audio is being captured. This is particularly useful in noisy environments where the model might otherwise interpret background sound as input.
The updates also extend well beyond the interface itself. Anthropic expanded language support in Voice Mode to cover approximately 18 additional languages, including Russian and German. The original Voice Mode beta, which launched in late May 2025, was English-only.
Where this fits in the competitive picture
The model selector is a differentiator worth noting. Giving users the ability to choose their model inside a voice session is something neither OpenAI’s consumer app nor Google’s Gemini app currently offers in quite the same way.
There has been no formal announcement or blog post from Anthropic outlining these updates, although users have reported experiencing the new features firsthand.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) ·