xAI Launches Grok Voice 2.0: #2 in Speech-to-Speech, 0.70s Latency
ArtificialAnlys · x · 2026-07-30
xAI has released Grok Voice Think Fast 2.0. According to Artificial Analysis, the model ranks #2 on the Speech-to-Speech Index (82.9%) and takes the #1 spot on the Tau Voice benchmark for agentic performance (56.5%).
Its major competitive edge is speed: it achieves an average Time to First Audio of just 0.70 seconds. It is the only model in the top five with latency under 1 second, significantly outpacing GPT-Realtime-2 High (1.14s) and Qwen Audio 3.0 (4.02s).
Priced at $4.80 per hour of input audio, it is more expensive than some peers but remains 2.2x cheaper than GPT-Realtime-2.1 High ($10.75).
More from Models
- Researchers Find Anomalous Narrative Fulfillment Tendencies in Claude Opus 5 Base Mode — repligate · 2026-07-30
- Specific Prompt Triggers Anomalous User-Completion Behavior in Claude Opus 5 — matthen2 · 2026-07-30
- New Technologies Like MLA and GRPO are Decentralizing AI Open Source — zephyr_z9 · 2026-07-30
- Users Complain About Claude Opus Being Too Preachy for a Personal Assistant — sidjustice_ · 2026-07-30
- Users Excited as Opus 5 is 'Cooking' with Impressive Capabilities — chrisfirst · 2026-07-30
- Claude Experiences Major Service Outage, Disrupting Access — Polymarket · 2026-07-30