Voicebench

Which speech model hears real people best, and fastest?

← All models

Voxtral Mini 4B Realtime

MistralApache-2.0streaming
Score
40/ 100
#19 of 21 models
Word errors
5.5%
Typical wait
0.82 s
slow: 0.96 s
Under a second
90%

How the score adds up

Accuracy
5.5% word errors · counts 60%
50
Typical wait
0.82 s typical wait · counts 25%
18
Slow wait
0.96 s for the slowest 1 in 10 · counts 15%
38

What it heard

39.5% word errors: 0 wrong, 68 missed, 0 added

good morning what is the weather like in london today and what about sunday set a timer for 10 minutes cancel that wake me up at 6 tomorrow send a message to david saying i am running about 15 minutes late so please start without me and i will join you soon yes send it remind me to buy eggs strawberries and orange juice when i leave work this evening play some quiet jazz in the kitchen a little louder please stop show me the fastest way to the airport from here and tell me whether there is heavy traffic on the bridge right now did anyone call while i was out schedule a meeting with emma for next wednesday at 3 and move my usual team lunch to thursday search for a vegetable soup recipe that serves 4 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

0.73 s0.81 s0.81 s0.78 s0.78 s0.82 s0.88 s0.91 s0.72 s0.16 s0.77 s0.50 s0.89 s–––––––
How it ran
Engine
llama-server (realtime)
Download
4.72 GB
Memory used
14.0 GB
Load time
4.7 s
Model files
sha256 aa1df815ca5f95c0… · sha256 f28ae3e40f2b610c…