Spider 2.0-Lite: leaderboard
547-problem text-to-SQL benchmark spanning BigQuery, Snowflake, and SQLite from real enterprise use cases, translating natural language into SQL using schema metadata.
Metric: Accuracy (%). Source: spider2-sql.github.io. Saturation forecast: Around January 2027. 37 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GLM-5.2 | 76.23 | #73 |
| 2 | Claude Opus 4.5 | 65.81 | #83 |
| 3 | O3 | 55.21 | #124 |
| 4 | GPT-5 | 54.66 | #98 |
| 5 | O4 Mini | 53.02 | #180 |
| 6 | DeepSeek R1 | 52.28 | #251 |
| 7 | DeepSeek V4 Flash | 42.4 | #137 |
| 8 | Claude Sonnet 4.5 | 41.86 | #142 |
| 9 | GLM-5 | 32.36 | #141 |
| 10 | O1 Preview | 30.35 | #193 |
| 11 | Claude 3.7 Sonnet (20250219) (Thinking) | 28.52 | #201 (Claude 3.7 Sonnet (20250219)) |
| 12 | Claude Sonnet 4 (20250514) | 27.79 | #213 |
| 13 | DeepSeek V3 | 26.14 | #321 |
| 14 | Claude 3.7 Sonnet (20250219) | 25.41 | #201 |
| 15 | O3 Mini (2025-01-31) | 23.4 | #240 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=spider-2-0-lite · How It Works · Data refreshed daily, snapshot 2026-10-20.