CoVoST2
Multilingual
audio
+
+
+
+
About
CoVoST2 is a large-scale multilingual speech-to-text translation benchmark covering translations from 21 languages to English and from English into 15 languages. This comprehensive benchmark evaluates AI models' ability to perform cross-lingual speech translation, testing both speech recognition and translation capabilities simultaneously. CoVoST2 provides essential evaluation for multilingual speech processing and cross-language communication systems.
+
+
+
+
Evaluation Stats
Total Models2
Organizations1
Verified Results0
Self-Reported2
+
+
+
+
Benchmark Details
Max Score1
Language
en
+
+
+
+
Performance Overview
Score distribution and top performers
Score Distribution
2 models
Top Score
39.2%
Average Score
38.8%
High Performers (80%+)
0Top Organizations
#1Google
2 models
38.8%
+
+
+
+
Leaderboard
2 models ranked by performance on CoVoST2
| License | Links | ||||
|---|---|---|---|---|---|
| Dec 1, 2024 | Proprietary | 39.2% | |||
| Feb 5, 2025 | Proprietary | 38.4% |
