← Back to dashboard

Model Performance

How major LLMs are improving over time.

Model explorer

Pick a company and model to review the dashboard benchmark rows we keep for construction work-area themes — chips show loaded suites only (no citations in this panel).

Company

Choose provider

Models (flagship-focused)

Selected model

GPT-5.2 Pro

Company: OpenAI

Released: Dec 2025

Dashboard overall score: 93.2%

Limited loaded benchmark coverage for this model.

Loaded benchmark suites

GPQA Diamond · no tools — 93.2%