Dograh
Experiments  /  TTS verbalization benchmark  /  kokoro-82m
Model · EXP-001

kokoro-82m

Ranked 1 of 7 on the TTS verbalization benchmark. Strongest on initialism or acronym, weakest on hashtag or mention.

At a glance
Said right61%share of all 1,080 sentences
Rank1 of 7among the seven models
Cost per clip0.31scomputing time per audio clip
Likely range58%–64%where the true score most likely sits
Strengths and weak spots

Where kokoro-82m does best and worst

Its five strongest and five weakest types of text, by the share of sentences it said right.

Best at
Initialism or Acronym
1.00
Musical Notation
1.00
Stock Ticker
1.00
Currency
0.97
Vehicle or Product Code
0.95
Breaks on
Hashtag or Mention
0.00
Geographic Coordinates
0.00
Fractions
0.10
Date
0.10
ISBN
0.15
A real miss

Hashtag or Mention: what kokoro-82m actually said

The text
Follow #GreenBreak for updates during the conference.
Should be said as
Follow hashtag green break for updates during the conference.
kokoro-82m said
"follow green break for updates during the conference"
All 27 types

Every type of text, best to worst

The share of sentences kokoro-82m said right on each of the 27 types. Each type has 40 test sentences.

Initialism or Acronym
1.00
Musical Notation
1.00
Stock Ticker
1.00
Currency
0.97
Vehicle or Product Code
0.95
Decimal
0.90
Ordinal
0.90
Roman Numeral
0.88
Abbreviation
0.85
Cardinal
0.80
Legal Reference
0.78
License Plate or Serial Numbers
0.72
Time
0.72
Mathematical Expression
0.68
Chemical Formula
0.62
Unit
0.62
Address
0.55
Sports score
0.53
Biological Classification
0.45
Phone Number
0.42
URL or Email
0.42
Version Numbers
0.30
ISBN
0.15
Date
0.10
Fractions
0.10
Geographic Coordinates
0.00
Hashtag or Mention
0.00