← Back to dashboard

Model Performance

Track how the leading AI systems are improving.

Compare models

Pick a company, then a model.

Choose company

Company

Models

Model

GPT-5.2 Pro

OpenAI·Released Dec 2025

93.2overall · /100
Single-category

Cost: $$$$

How to read this model
  • Overall — Average of the category scores below (0–100).
  • Coverage (Single-category) Based on one category — not a full picture of overall strength. (One work category)
  • Cost Rough relative tier — not a vendor price.
  • Construction fit — The chips below tilt toward where this snapshot's scores are strongest; always verify on your own jobs and tools.

Scores by category

  • GPQA Diamond (no tools)Reasoning93.2%

Best construction-office fit

  • Premium reasoning runs
  • Deep technical reviews
  • High-stakes drafting with oversight