Google Research announced its new system, Gemini-SQL2, which has the ability to create SQL queries from natural language. Using the Gemini 3.1 Pro infrastructure, this technology automates complex knowledge base queries, eliminating the need for users to write manual code. The system, which reached a success rate of 80.04% in the tests carried out on the BIRD data set, is not satisfied with just grammatical accuracy, but also guarantees that the queries work without errors in real knowledge base environments. This development aims to speed up business processes by allowing users, especially those with low technical knowledge, to perform in-depth analysis on large data sets.

Gemini-SQL2 Stands Out in the Section with Its Performance

Traditional text-to-SQL systems often measure theoretical accuracy by focusing on the structure of the query. However, this new model developed by Google checks whether the query gives the correct result by running it personally, as in the BIRD benchmark tests.

This approach prevents errors encountered in real-world applications.

This new technology greatly reduces the need for manual software in data analysis processes.

Database Administration Becomes More Accessible

The 80.04% success rate offered by Gemini-SQL2 proves that the model accurately understands even multi-layered and complex database questions. This level of automation allows analysts and business development experts to quickly access data without encountering technical barriers. For companies, this means democratizing data-based decision-making processes.

The accuracy provided by the system has a great time-saving potential for institutions working with high volumes of information. Replacing the coding stage with natural language commands reduces software development costs and increases operational efficiency.

Artificial intelligence models automatically manage complex information retrieval processes on behalf of the user.

Future Data Analytics Transformation Expected

This move by Google Research once again reveals the transformative effect of productive artificial intelligence on database management. We are entering a period in the future where even employees who do not know the SQL language will be able to produce meaningful reports from the most complex knowledge bases. Gemini-SQL2 is positioned as one of the most powerful tools of this transformation.

Do you think SQL queries created by artificial intelligence can completely replace data analysts? Do not forget to share your opinions and experiences on this subject with us in the comments section.