AIMultiple - Text-to-SQL (Strict Execution Match): leaderboard
AIMultiple Text-to-SQL, from the same BIRD multi-database routing runs: the model's final SQL is executed and its result set compared with BIRD's gold query, without the domain hint, over the questions the model routed correctly, so denominators differ by model. Strict execution match; %, higher is better.
Metric: Execution match against BIRD gold (%, over correctly routed questions). Source: aimultiple.com. Saturation forecast: Around 2030. 81 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 | 74.9 | #3 |
| 2 | Gemini 4 Argon | 64.3 | #9 |
| 3 | Gemini 3.8 Flash | 63.5 | #30 |
| 4 | Gemini 3.6 Flash | 61.4 | #57 |
| 5 | Gemini 3.7 Flash | 59.4 | #24 |
| 6 | Claude Sonnet 5.5 | 59.4 | #14 |
| 7 | Gemini 3 Flash (Preview) | 55.1 | #90 |
| 8 | Claude Fable 5.1 | 54.7 | #10 |
| 9 | Gemini 3.1 Pro (Preview) | 53.6 | #64 |
| 10 | Claude Fable 5 | 53.1 | #20 |
| 11 | GPT-6 Astra | 52.8 | #6 |
| 12 | GPT-6.1 Sol | 52.4 | #7 |
| 13 | Claude Opus 5 | 52.3 | #16 |
| 14 | GPT-6.1 Pro Sol | 51.3 | #11 |
| 15 | Grok 4.6 | 50.4 | #29 |
Interactive version: theaggregate.ai/benchmark?slug=aimultiple-text-to-sql-strict-execution-match · How It Works · Data refreshed daily, snapshot 2026-11-05.