*This dataset supports online use.Click here to jump.

HTSC-2025 is a benchmark dataset for predicting the critical temperature of normal-pressure high-temperature superconductors, released by Renmin University of China and Hefei National Laboratory in 2025. The related paper results are "HTSC-2025: A Benchmark Dataset of Ambient-Pressure High-Temperature Superconductors for AI-Driven Critical Temperature Prediction", which aims to provide standardized and comparable test samples for the model to promote the progress and verification of superconductor prediction tasks.

This dataset contains approximately 140 materials, stored in JSON/Parquet format for easy processing. It emphasizes verifiable predictions based on electron-phonon coupling at ambient pressure and covers a wide range of structural systems. It is suitable for Tc prediction and evaluation of machine learning models, as well as for materials screening and characterization studies. It serves as an important benchmark resource for advancing superconducting material discovery.

Data composition:

• The chemical formula and structural information of the material (usually described in CIF format crystal structure).
• Theoretical predictions of the critical temperature Tc (from first-principles calculations + BCS-type mechanisms such as phonon-electron coupling).
• The material family/structural system label, for example: X2YH6X_2YH_6X2​YH6​, MXH₃ type perovskite, M₃XH₈ type, LaH₁₀ type cage-like, two-dimensional honeycomb structure, etc.