Skip to content
trending

Benchmarks

VUI Labs models, access and benchmark guide

VUI Labs model and access guide with a practical US-English evaluation brief and links to publisher materials.

By 1 min read
VUI Labs models, access and benchmark guide

👁 111 reads

Luna TTS should be evaluated by the voice VUI Labs actually serves. The project describes the model and the developer site documents endpoint usage. On a call desk I would grade address intelligibility at phone bandwidth before listening for studio polish.

VUI Labs describes a Luna family that can generate a system voice or reuse a custom voice made from a short reference. Its documentation also describes streaming audio while it is generated. The underlying research covers Chinese, English, Japanese and Korean speech. Those capabilities raise two different questions for a US call desk. Can the first words play promptly on the target connection, and can a caller still distinguish “14 B” from “40 D”? For the second question I would use the telephone recording as the listening sample. The family paper describes the architecture, while the provider’s documentation establishes which controls are exposed to an account.

Choosing the first job

Have listeners transcribe a fictional US apartment address from a telephone-bandwidth rendering.

Models in the November 17 snapshot

Model and track Placement Reported score
Luna TTS (AA provider voice) 21 1196 Elo

Release record. Publisher documentation identifies the product family. The ranked configurations include Luna TTS.

Google · Your sourcesQuantslant higher in your Google - one clickAdd to preferred sources →

Sources

Kinetyra

// read next

More on this from QUANTSLANT

Piotr Olszewski

Editorial team

Independent AI publication. Models and tools are tested on real tasks, prices are checked against official rate cards, and articles are updated when things change.

read nextVoice AI agents ranked for live conversationai toolsVoice AI agents ranked for live conversationCompare speech-to-speech systems on conversational preference, then audit interruptions, corrections and confirmations.

Most read

  1. 1Best AI Video Generator Apps for Your Phone (2026)
  2. 2Claude Opus 5.5 Motion Graphics: 60s Showreel from One Prompt
  3. 3Wan 2.2 on RTX 4090: timings, VRAM and four test clips
// site map
Quantslant in numbers 1,469posts in archive39news from the last 7 days18video models in the observatory
🎬 AI Video Benchmarks: Higgsfield DoP vs OpenAI Sora • Higgsfield vs Runway Gen-3 • AI Video Credit Calculator
× ‹ enlarge ›