How to find an apprenticeship?

We provide an official service to search through available apprenticeships. To get started, create an account here, specify the desired region, and your preferences. You will be able to search through all officially registered open apprenticeships.

You can contact the apprenticeship office through our official phone hotline above, or with the web-form below. We generally respond to written requests within 7-10 days.

// VoXClash

Blind pairwise evaluation across languages.

Model-vs-model Elo rankings, human reference anchors, and per-language quality metrics for state-of-the-art TTS systems.

Hindi

English

Telugu

Model leaderboard

Hindi · Elo from model-vs-model votes only

Head-to-Head win-rates

Row model's win % against column model

Compare-to-human voting is coming soon

Model-vs-model rankings are live above. Expert human recordings will be integrated as benchmark baselines soon.

Latency

P50 is the usual wait. P25–P75 is how jumpy that wait is. Min–max is best vs worst. Silence is hush at the start of the clip.

Issue tags by model

Tagged issues as a share of each model's model-vs-model matchups

Model leaderboard

English (IN) · Elo from model-vs-model votes only

Head-to-Head win-rates

Row model's win % against column model

Compare-to-human voting is coming soon

Model-vs-model rankings are live above. Expert human recordings will be integrated as benchmark baselines soon.

Latency

P50 is the usual wait. P25–P75 is how jumpy that wait is. Min–max is best vs worst. Silence is hush at the start of the clip.

Issue tags by model

Tagged issues as a share of each model's model-vs-model matchups

Model leaderboard

Telugu · Elo from model-vs-model votes only

Head-to-Head win-rates

Row model's win % against column model

Compare-to-human voting is coming soon

Model-vs-model rankings are live above. Expert human recordings will be integrated as benchmark baselines soon.

Latency

P50 is the usual wait. P25–P75 is how jumpy that wait is. Min–max is best vs worst. Silence is hush at the start of the clip.

Issue tags by model

Tagged issues as a share of each model's model-vs-model matchups

Frequently Asked Questions

Rankings come from blind listening tests: people hear two clips of the same script and pick which sounds better. Scores are Elo ratings from those pairwise choices, and not automated text or speech metrics.

Higher Elo means the model wins more head-to-head comparisons. The bracket is a 95% uncertainty range: with current votes, the score could reasonably sit within that band.

Latency measures how quickly audio starts (time-to-first-audio). It does not affect quality rank, a model can be fast but sound worse, or slow but sound better.

This doesn’t have to end here

Accuracy is Intelligence

Book a Demo