Sapienz Receives ISSTA 2026 Impact Paper Award
Meta's Sapienz now uses LLM agents to test functional correctness beyond crash detection
“Beyond detecting crashes, large language model (LLM) agents can now explore apps and reason about whether their behavior matches what users should expect, extending automated testing from reliability to functional correctness and, increasingly, automated repair.”
Meta's Sapienz automated Android testing system, which runs millions of daily tests across Meta platforms, received the ISSTA 2026 Most Influential Paper Award for its decade of impact since a 2016 UCL research paper. The more forward-looking signal is that Sapienz has evolved to integrate LLM agents capable of reasoning about functional correctness — not just crashes — and is moving toward automated repair. This marks a meaningful step in applying LLMs to large-scale production software testing infrastructure.