IndicDB (Zero-Shot with Evidence) - Hindi: leaderboard
Metric: Execution accuracy (%) of generated SQL on the Hindi variant of the IndicDB Text-to-SQL tasks (20 databases built from Indian government open data, join depth up to six), zero-shot prompting with the task's external evidence; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 4 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V3.2 | 67.73 |
| 2 | MiniMax-M2.7 | 67.13 |
| 3 | Llama 3.3 70B Instruct | 61.27 |
| 4 | Qwen 3 8B | 42.44 |
Interactive version: theaggregate.ai/benchmark?slug=indicdb-zero-shot-with-evidence-hindi · How It Works · Data refreshed daily, snapshot 2026-10-07.