TNODEV: Toolbox for Neural ODE Verification

TNODEV has been introduced as the first sound formal verifier for neural ordinary differential equations (neural ODE), addressing the need for formal verification in safety-critical applications. It features a falsification checker, a fast interval-based reachability backend utilizing continuous-time mixed monotonicity, and a verification loop with input-set splitting heuristics, enabling iterative refinement. TNODEV supports verification for various neural ODE configurations and has been benchmarked against existing tools like NNV~2.0 and CORA, demonstrating its capabilities in safe-set inclusion and classification robustness, which is crucial for practitioners developing reliable AI systems.