Skip to content

Benchmark

FLEURS-35 · Realtime macro WER/CER

Scores are percentages. Lower is better.

Models measured
2
Scores recorded
5
Sources
1
IQ weight
No weight published

Score distribution

5.9%
Mean
6%
Median
1.1%
Std dev
4%
10th percentile
7.4%
90th percentile
4%
Min
4%7.4%
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1Voxtral Mini Transcribe 3Voxtral 3 official release chart · offline Mini Transcribe 3 baseline on same 35-language set· Reported by Mistral AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.4%
2Voxtral Mini Transcribe Realtime 3.0Voxtral 3 official release chart · realtime · 2,400 ms latency configuration· Reported by Mistral AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.5.7%
3Voxtral Mini Transcribe Realtime 3.0Voxtral 3 official release chart · realtime · 960 ms latency configuration· Reported by Mistral AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.6%
4Voxtral Mini Transcribe Realtime 3.0Voxtral 3 official release chart · realtime · 480 ms latency configuration· Reported by Mistral AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.6.4%
5Voxtral Mini Transcribe Realtime 3.0Voxtral 3 official release chart · realtime · 240 ms latency configuration· Reported by Mistral AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.7.4%

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.