ResearchBenchmarksMulti-task Driving Scene Understanding Robustness on The Dolphins (Lvl. 3)Follow44.59Final ScoreNutVLM35.448437.821740.19542.5683Feb 9, 2026Evaluation ResultsMethodMethodLinksFinal ScoreNutVLM2026.0244.59AAA2026.0243.95MS2026.0243.7Scene–CADA2026.0243.56NRP2026.0243.45Bit-Red2026.0243.2JPEG2026.0241.1TVM2026.0235.8