MuLA-Bench - Thai: leaderboard
The Thai questions of MuLA-Bench (Yang et al., 2026), a benchmark of open-ended questions over long in-the-wild recordings, answered from the audio alone, covering spoken content and natural sounds in original Thai recordings rather than translations. Each answer is judged correct or wrong by GPT-5.6 Sol against the reference with rubric and timing checks; accuracy in percent, higher is better.
Metric: Accuracy (%; the 309 Thai-language questions over both evidence tracks; open-ended questions over in-the-wild long-form recordings (median 36.8 minutes, capped at three hours) in the source language, answered from the full recording without transcripts, audio clipped only to a local model's native input limit; each answer judged correct or not by GPT-5.6 Sol on a route-specific binary rubric with programmatic structural and 3-second timing checks; missing or invalid answers scored wrong). Source: arxiv.org. Saturation forecast: Around February 2027. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.8 Flash | 72.5 | #36 |
| 2 | Gemini 3.7 Flash | 72.2 | #32 |
| 3 | Gemini 3.1 Pro (Preview) | 58.6 | #91 |
| 4 | Qwen 3.5 Omni Plus | 57.9 | #237 |
| 5 | Qwen3.5 Omni Flash | 43 | #548 |
| 6 | Muse Spark 1.2 | 31.1 | #97 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=mula-bench-thai · How It Works · Data refreshed daily, snapshot 2026-11-12.