We are proud to share that our proprietary AI system has been recognised globally on the Spider 2.0 Snow leaderboard, a leading benchmark for real world Text-to-SQL and database reasoning. This recognition highlights our system’s ability to translate natural language into complex database queries, a core AI capability used every day by developers, tech ops, and analysts to power analytics and decision making at scale.
Being featured on the global capability leaderboard of the Spider 2.0 benchmark marks an important milestone for our engineering teams and reinforces our focus on building AI software designed for real world use. We are also the first Indian company to feature on this global leaderboard, placing India firmly on the map for advanced AI software development.
What Spider 2.0 measures and why it matters
Spider 2.0 is an industry recognised benchmark presented at ICLR 2025. It is designed to evaluate how effectively AI systems convert natural language questions into accurate SQL queries over complex, production grade databases.
Unlike academic benchmarks, Spider 2.0 mirrors how AI systems are actually used by developers, tech ops, and analysts in real enterprise environments. The benchmark tests advanced reasoning across complex schemas, multi-table joins, constraints, nested logic, and real business rules that are common in everyday analytics and operational workflows.
Our global ranking on Spider 2.0
We achieved a score of 82.63 on the Spider 2.0 benchmark, ranking second globally and outperforming several leading international research institutions and global technology companies.
This performance reflects our strength in building AI systems that are accurate, reliable, and consistent when operating on large, complex, real world databases. It also reinforces our belief that AI software should be evaluated based on how well it performs in production environments.
Prism and our approach to Text-to-SQL intelligence
Our performance on Spider 2.0 was powered by Prism, our proprietary in-house AI software system for advanced Text-to-SQL reasoning. Prism enables natural language queries to be translated into accurate SQL outputs across large and complex databases, helping teams extract insights faster and with greater confidence.
Prism is built on our self-organising swarm architecture, where multiple specialised AI agents collaborate to plan, generate, validate, and refine SQL queries. This multi-agent design improves accuracy, reliability, and consistency by continuously validating logic, constraints, and query structure before execution.
We have designed Prism to operate on large, mission critical databases managed by developers, tech ops, and analysts, reflecting our focus on building AI systems that work reliably in real production environments.
Building AI software that works in the real world
“This milestone validates our AI software engineering capabilities and reflects the depth of our proprietary technology. Being the only Indian company on the global capability leaderboard of the Spider 2.0 benchmark reinforces our belief that world class AI software can be built in India and used at scale by developers, tech ops, and analysts,” said a Paytm spokesperson.
For us, this recognition is not just about rankings. It is about building AI systems that solve real problems, operate at scale, and can be trusted in everyday workflows.
The people behind the work
This achievement was driven by the efforts of Soham Acharya, Senior Manager of Data Science, and Anshul Chauhan, Lead Software Engineer. Their work played a key role in advancing AI research and applied AI systems and reflects the depth of engineering talent behind our AI platforms.
What this means going forward
As AI becomes deeply embedded in how organisations interact with data, natural language access to complex databases will become a foundational capability. Our recognition on Spider 2.0 reinforces our long-term commitment to building reliable, scalable, and production-ready AI systems with real world impact.
We will continue to invest in AI research, strong engineering practices, and applied innovation as we build technology that works at scale.
Building the next AI frontier from India
Just as Paytm helped transform how India pays, we are now building AI systems designed for real-world scale. Engineered in India, by Indians, and trusted globally.
Come build with us. Explore career opportunities at Paytm and be part of our growing AI journey, from India to the world. Apply today.
Author
