AudioBench - CoVoST2 EN-ID BLEU: leaderboard
AudioBench CoVoST2 EN-ID BLEU speech-translation BLEU slice measuring translation quality between the named source and target languages.
Metric: Mean Score (%). Source: github.com. Saturation forecast: Estimated already saturated. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | SeaLLMs-Audio-7B | 27.88 | #1018 |
| 2 | Qwen2-Audio-7B-Instruct | 16.33 | #880 |
| 3 | Phi-4 Multimodal Instruct | 15.01 | #906 |
| 4 | SALMONN-7B | 14.1 | #1293 |
| 5 | Qwen-Audio-Chat | 4.1 | #1244 |
| 6 | Whisper Large-v3 | 1.6 | #1271 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=audiobench-covost2-en-id-bleu · How It Works · Data refreshed daily, snapshot 2026-10-20.