BIRD-Interact (c-Interact): leaderboard
Interactive text-to-SQL evaluation where models engage in conversational multi-turn interactions with databases to resolve user queries.
Metric: Normalized Reward. Source: github.com. Saturation forecast: Around December 2026. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 20.92 | #152 |
| 2 | O3 Mini | 20.27 | #272 |
| 3 | Claude Sonnet 4 | 18.35 | #199 |
| 4 | Qwen 3 Coder 480B A35B | 17.75 | #279 |
| 5 | DeepSeek V3.1 | 15.15 | #264 |
| 6 | Claude 3.7 Sonnet | 13.87 | #249 |
| 7 | GPT-5 | 12.58 | #98 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=bird-interact-c-interact · How It Works · Data refreshed daily, snapshot 2026-10-20.