Google Research's Gemini-SQL2 tops text-to-SQL benchmarks by a wide margin

Google Research announced the release of Gemini-SQL2, a text-to-SQL capability powered by the Gemini 3.1 Pro model, which achieved an execution accuracy of 80.04% on the BIRD single-model leaderboard. This score reflects the model's effectiveness in converting natural language queries into SQL commands, highlighting its potential for practical applications in database querying. The implementation pattern discussed suggests a schema-grounded approach, which could enhance the reliability of SQL generation in various use cases.