List your company

Pairwise Solution Preference Prediction on Data-centric Solution Preference Dataset (test)

61.5Average AccuracyDeepSeek-V3.2 (Thinking mode)
49.5452.64555.7558.855Jan 9, 2026

Evaluation results

MethodLinks
DeepSeek-V3.2 (Thinking mode)Thinking mode=true, Temperature=1, Reasoning instructions=Wei et al., 2023; Kojima et al., 20232026.0161.559.366.957.460.463.95758.962.166.8
GPT-5.1Version=gpt-5.1-2025-11-13, Temperature=1, Reasoning instructions=Wei et al., 2023; Kojima et al., 20232026.0158.855.46357.455.561.659.757.259.262.2
Complexity Heuristicdimensions=Code Engineering, Model Architecture, Data Pipeline2026.0150.8—————————
Random Guess2026.0150—————————