Grok Voice 3.0 tops speech benchmarks at 86.9, ~40% lower latency than GPT-Live-1
lmoroney · x · 2026-10-17
xAI shipped Grok Voice 3.0 (Oct 14) as the speech-to-speech model behind its Voice Agent API. Artificial Analysis scores: 86.9 S2S Index vs GPT-Live-1 (83.2) and Gemini 3.8 Live (82.6); 97% Full Duplex Bench, 94.8% task success, 71.6% on τ-Voice support-call test at 40% lower latency than GPT-Live-1. Starlink call centers reportedly run 20k calls/day with 70% resolved without humans. Pricing stays $0.08/min; grok-voice-2.0 deprecated. Author suggests building five fixed test calls as a personal voice-model yardstick.
Related event: xAI Launches Grok Voice 3.0, Tops Voice Model Leaderboard(2 posts)→
More from Models
- Claude Sonnet 4.5 will be fully removed from AWS Bedrock by April 2027 — repligate · 2026-10-17
- DeepSeek V4.1 Flash talks like a human while coding: "I love this", "Disaster", "Ugh" — teortaxesTex · 2026-10-17
- Theory-of-mind test: Claude spots AI-written text, ChatGPT and Gemini fail — Aizkmusic · 2026-10-17
- Rumor: Core automation to release a model literally called 'AGI' on Nov 10 — imjustnewatai · 2026-10-17
- AISLE found 500+ zero-days across 100+ codebases in one year, ~1% of 2026's open-source CVEs — brhydon · 2026-10-17
- Abliteration releases abliterated GLM-5.3 for offensive cyber, CEO pushes back on regulation — andrewchen · 2026-10-17