AudioBench - IMDA AR Sentence: leaderboard
AudioBench IMDA AR Sentence slice measuring audio reasoning and recognition from spoken inputs.
Metric: Mean Score (%). Source: github.com. Saturation forecast: Around 2029. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen-Audio-Chat | 3.93 | #1244 |
| 2 | Qwen2-Audio-7B-Instruct | 3.61 | #880 |
| 3 | SeaLLMs-Audio-7B | 3.58 | #1018 |
| 4 | Phi-4 Multimodal Instruct | 3.52 | #906 |
| 5 | SALMONN-7B | 2.52 | #1293 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=audiobench-imda-ar-sentence · How It Works · Data refreshed daily, snapshot 2026-10-20.