AssemblyAI's Universal 3.6 Pro tops STT benchmark with 1.77% WER and 91ms median latency
AssemblyAI · x · 2026-09-30
AssemblyAI launched Universal 3.6 Pro, which posts the lowest word error rate (1.77% on 1,000 Pipecat clips, down from 1.93%) among 16 models in Cekura's speech-to-text benchmark. Final-text latency halved: median 91ms (from 180ms) and P95 232ms (from 692ms), beating Google Chirp 3 and others on accuracy and speed.
Related event: AssemblyAI's Universal 3.6 Pro Tops Realtime Speech Benchmarks(3 posts)→
More from Models
- 32 researchers spent 8 months on the most comprehensive survey of tokenization ever — christopher · 2026-10-01
- GPT-6.1 Sol reportedly tops MathArena: 86.3% accuracy at $0.94 vs Astra's 81.9% at $2.26 — 141_1337 · 2026-10-01
- Defending Astra's Caution: An Agent That Stops to Ask Is Doing It Right — brandon_galang · 2026-10-01
- Dev complains Opus 5.5 burned 80% of quota in 3 days, switching models — MaziyarPanahi · 2026-10-01
- Liquid AI's LongevityBench: compact LFMs beat frontier models on several aging tasks — JosephJacks_ · 2026-10-01
- GPT 6.1 Astra ultra code mode builds a three.js ocean scene — OpenAIDevs · 2026-09-30