VakQA (IndicWhisper ASR Transcripts): leaderboard
Metric: Judge score (1-5) (2,001 Telugu spoken factoid questions from quiz-style interactions across six domains, with human-verified reference answers; each answer scored 1 to 5 for correctness against the reference by Gemini-2.5-Flash as judge (the best-aligned judge in the human study); mean over the questions; input: the Telugu transcript from IndicWhisper ASR). Source: arxiv.org. Saturation forecast: Estimated already saturated. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Flash | 3.09 |
| 2 | Gemma 3 27B | 2.22 |
| 3 | Gemma 3 12B | 1.77 |
| 4 | Gemma 3 4B | 1.33 |
Interactive version: theaggregate.ai/benchmark?slug=vakqa-indicwhisper-asr-transcripts · How It Works · Data refreshed daily, snapshot 2026-09-26.