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

Google Research has introduced Gemini-SQL2, an specialized model built on Gemini 3.1 Pro that achieves 80.04 percent accuracy on the BIRD benchmark for text-to-SQL translation. This performance surpasses current top models from OpenAI and Anthropic, signaling a significant improvement in the ability of AI to convert natural language into functional database queries. By lowering the technical barrier for data interaction, this advancement is expected to enhance natural language integration across Google's suite of enterprise data services.