CASU - Contextual Reasoning: leaderboard
Metric: Accuracy (%; four-option multiple-choice questions on semi-synthetic auditory scenes that combine speech, sound events and background; contextual reasoning: reconcile a spoken claim with the events and background that support or contradict it). Source: arxiv.org. Saturation forecast: Around December 2026. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o Audio | 74.02 |
| 2 | Gemini 2.0 Flash | 73.3 |
| 3 | Qwen3 Omni 30B A3B Instruct | 71.18 |
| 4 | Qwen2.5-Omni-7B | 62.06 |
| 5 | Voxtral-Small-24B-2507 | 52.52 |
| 6 | Qwen2-Audio-7B-Instruct | 44.02 |
| 7 | SALMONN-13B | 43.21 |
Interactive version: theaggregate.ai/benchmark?slug=casu-contextual-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-26.