Humanness Index™ · Baseline reference

Human

humannessindex.vapi.ai

Human is the reference point on the Humanness Index, not a text to speech model.

Best ranked model: #0 Homo Sapien
Humanness: 100

Standings as of Jul 8, 2026, 10:19 UTC

Models on the Index: 1
Languages: —
Price / 1M chars: —

Human models on the Humanness Index™

Rank	Model	Humanness	Latency	Languages	Price / 1M chars
Baseline	Homo Sapien	100	—	—	—

Compare against the full Humanness Index™ rankings

About the Human baseline

Human is the reference point on the Humanness Index, not a text to speech model. The four original source voices on the Index (Clara, Emma, Godfrey, and Nelliot) are real people, and each of them recorded the same 20 customer support lines that every TTS model reads. Those recordings are the Human baseline.

Why it anchors the scale

In the arena a Human recording goes head to head against a cloned TTS version of the same voice reading the same line, so the question is literally which one is human. Human is pinned to a Humanness score of 100, and every model is then read as a share of that human mark. Because it is a reference and not a competitor, the Human baseline sits outside the model and provider counts and outside the latency comparison.

Human stats

Price / 1M chars: —No published pricing to cite yet.

Frequently asked questions

Is the Human baseline a text to speech model?: No. It is the four source voices, recorded by the real people behind them reading the same 20 lines the models read. It exists to anchor the top of the scale, so a model score reads as how close that model gets to a real person.
How is the Human baseline recorded?: The same four actors read all 20 customer support lines exactly as written, in a quiet room at a consistent level. The recordings are loudness normalized and converted to MP3 on ingest, the same handling the synthetic clips get, so the only difference a listener hears is the voice itself.

Other providers on the Index

ElevenLabsBest ranked model #3 · Eleven v3 CartesiaBest ranked model #14 · Sonic 3.5 xAIBest ranked model #1 · Grok TTS MiniMaxBest ranked model #2 · Speech 2.5 GradiumBest ranked model #20 · Gradium TTS Canopy LabsBest ranked model #4 · Orpheus InworldBest ranked model #7 · TTS-1.5-max Smallest.aiBest ranked model #18 · Lightning v3.1 NeuphonicBest ranked model #19 · neu_hq

Back to the Humanness Index™

Find the most human-sounding voice for your agent.

Compare the models in blind tests, read the methodology, or get in touch.

Read the methodology Star on GitHub

Build a TTS model? Add yours to the Index.

Rank

Model

Humanness

Latency

Languages

Price / 1M chars

Baseline

Homo Sapien

100

—

About the Human baseline

Why it anchors the scale

Frequently asked questions

Is the Human baseline a text to speech model?

No. It is the four source voices, recorded by the real people behind them reading the same 20 lines the models read. It exists to anchor the top of the scale, so a model score reads as how close that model gets to a real person.

How is the Human baseline recorded?

The same four actors read all 20 customer support lines exactly as written, in a quiet room at a consistent level. The recordings are loudness normalized and converted to MP3 on ingest, the same handling the synthetic clips get, so the only difference a listener hears is the voice itself.