gpt.college
A small campus for AI curiosity

A clearer place to explore AI.

Compare model performance, keep an eye on subscription paths, and play the browser games people make with AI — without turning every page into a dashboard.

Explore rankings Or go play something
11 models 25 benchmarks Updated 2026-06-12
AI Models Live

See what the evidence says.

Rankings, capability maps, benchmark context, model profiles, and comparisons.

Browse performance
Deals Growing

Compare where you subscribe.

Web, App Store, and Play Store pricing tracked as dated, sourced snapshots.

Open deals
Games Playground

Play the weird little things people made.

Community-made browser games, creator pages, comments, and one-click play.

Enter the arcade
Live performance snapshot

Start with the useful part.

Full rankings
Overall ranking 2026-06-12
  1. 01 Gemini 3.1 ProGoogle 73.6
  2. 02 Gemini 3 ProGoogle 69.2
  3. 03 OpenAI GPT-5.2 ThinkingOpenAI 68.9
  4. 04 OpenAI GPT-5.4OpenAI 66.4
  5. 05 DeepSeek-R1DeepSeek 59.1

A rollup of sourced benchmark rows. Coverage and methodology still matter.

Compare by capability
CodingAgenticReasoningMathKnowledgeResearch
Recently tracked All models
Gemini 3.1 Pro Google OpenAI GPT-5.4 OpenAI OpenAI GPT-5.2 Thinking OpenAI Claude Opus 4.6 Anthropic
Read the receipts

Sources stay close to the claims.

Model and benchmark pages keep sources, dates, coverage, and methodology visible so you can see how much confidence to place in a number.

Methodology → Benchmark library → About the site →