xAI Launches grok-voice-think-fast-1.0: Topping τ-voice Bench at 67.3%, Outperforming Gemini, GPT Realtime, and More

MarkTechPost / 4/26/2026

📰 NewsIndustry & Market MovesModels & Research

共有:

Key Points

xAI has launched its flagship voice model, grok-voice-think-fast-1.0, positioning it as a top-tier system for voice-based tasks.
The model reportedly achieves 67.3% on the τ-voice benchmark, which tops the benchmark leaderboards.
It outperforms multiple competitors, including Gemini and GPT Realtime, as well as xAI’s prior version across several real-world workflow categories.
Reported performance gains span retail, airline, and telecom use cases, suggesting broad applicability for enterprise voice assistants.
Overall, the release signals xAI’s continued push to improve voice intelligence quality and competitive parity with leading AI speech models.

The new flagship voice model outperforms Gemini, GPT Realtime, and its own predecessor across retail, airline, and telecom workflows