IndicDB (DIN-SQL with Evidence) - Bengali: leaderboard
Metric: Execution accuracy (%) of generated SQL on the Bengali variant of the IndicDB Text-to-SQL tasks (20 databases built from Indian government open data, join depth up to six), DIN-SQL decomposed prompting with the external evidence; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 4 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V3.2 | 65.36 |
| 2 | MiniMax-M2.7 | 62.73 |
| 3 | Llama 3.3 70B Instruct | 62.21 |
| 4 | Qwen 3 8B | 51.14 |
Interactive version: theaggregate.ai/benchmark?slug=indicdb-din-sql-with-evidence-bengali · How It Works · Data refreshed daily, snapshot 2026-10-07.