FrontierThe story, in brief
xAI Launches grok-voice-think-fast-1.0: Topping τ-voice Bench at 67.3%, Outperforming Gemini, GPT Realtime, and More
67.3%. That's xAI's new voice model on τ-voice Bench—beating Gemini and GPT Realtime across real workflows.
The KeyNews take
Why it matters
xAI's grok-voice-think-fast-1.0 sets a new benchmark for voice AI performance in enterprise use cases (retail, airline, telecom), signaling intensifying competition in multimodal reasoning models where voice quality directly impacts customer-facing deployments.
The key facts
5 to knowgrok-voice-think-fast-1.0 scores 67.3% on τ-voice Bench
Outperforms Gemini and GPT Realtime on same benchmark
Demonstrated across retail, airline, and telecom workflows
Outperforms own predecessor model
Published April 25, 2026 on MarkTechPost
Go to the source
MarkTechPostmarktechpost.com
Publisher excerpt: The new flagship voice model outperforms Gemini, GPT Realtime, and its own predecessor across retail, airline, and telecom workflows