Voicebench

Which speech model hears real people best, and fastest?

← All models

VibeVoice-ASR-Streaming 1.5B

MicrosoftMITwhole utterance
Score
20/ 100
#21 of 21 models
Word errors
39.2%
Typical wait
0.32 s
slow: 0.51 s
Under a second
100%

How the score adds up

Accuracy
39.2% word errors · counts 60%
0
Typical wait
0.32 s typical wait · counts 25%
45
Slow wait
0.51 s for the slowest 1 in 10 · counts 15%
59

What it heard

32.6% word errors: 42 wrong, 11 missed, 3 added

good morning what is the it weather like to in live london in today london and hola what qu about tal sunday set a timer for 10 minutes can cancel you that hear wake me looking up for at 6 months tomorrow my send name a is message mr to david tseng saying i am running about 15 minutes late so please start with without that me and i will join you soon what yes is send that it remind me to buy eggs strawberries and orange juice when i leave work this evening there is play some quiet jazz in the kitchen a little louder please stop tell show me the fastest way to the airport from here and tell me what whether it there is heavy traffic on the bridge right now did anyone call while i was out good schedule morning a it meeting is with anna emma for next wednesday at 3 i am and in move my usual team most lunch of to the thursday it search is for a vegetable soup recipe that serves for 1 4 yeah no the other 1 how do you say thank you in spanish can you explain how a rainbow forms using simple words that a young child could easily understand thanks that is all

0.28 s0.34 s0.30 s0.31 s0.30 s0.30 s0.51 s0.28 s0.43 s0.33 s0.30 s0.27 s0.50 s0.33 s0.46 s0.35 s0.30 s0.33 s0.43 s0.25 s
How it ran
Engine
crispasr
Download
1.86 GB
Memory used
1.9 GB
Load time
0.6 s
Model files
sha256 bc81d02e195dd248…