DeepTest Tool Competition 2026: Benchmarking an LLM-Based Automotive Assistant
For researchers and practitioners in LLM testing, this competition provides a benchmark and initial comparison of automated testing tools for LLM-based applications.
The paper presents the first LLM Testing competition at DeepTest 2026, where four tools benchmarked an LLM-based car manual assistant to find inputs causing it to miss warnings. The competition evaluated effectiveness and diversity of failure-revealing tests.
This report summarizes the results of the first edition of the Large Language Model (LLM) Testing competition, held as part of the DeepTest workshop at ICSE 2026. Four tools competed in benchmarking an LLM-based car manual information retrieval application, with the objective of identifying user inputs for which the system fails to appropriately mention warnings contained in the manual. The testing solutions were evaluated based on their effectiveness in exposing failures and the diversity of the discovered failure-revealing tests. We report on the experimental methodology, the competitors, and the results.