NALA_MAINZ at BLP-2025 Task 2: A Multi-agent Approach for Bangla Instruction to Python Code Generation

arXiv — cs.CLMonday, November 24, 2025 at 5:00:00 AM
  • JGU Mainz's NALA_MAINZ team has secured first place in the BLP-2025 Shared Task on Code Generation from Bangla Instructions, achieving a $Pass@1$ score of 95.4 with their innovative multi-agent-based pipeline. This system includes a code-generation agent that creates an initial solution, followed by a debugger agent that refines the code based on test failures.
  • This achievement highlights the effectiveness of multi-agent systems in AI-driven code generation, showcasing the potential for improved programming tools and methodologies, particularly for non-English programming languages like Bangla.
— via World Pulse Now AI Editorial System

Was this article worth reading? Share it

Recommended apps based on your readingExplore all apps

Ready to build your own newsroom?

Subscribe to unlock a personalised feed, podcasts, newsletters, and notifications tailored to the topics you actually care about