A newly launched platform called AI IQ, developed by Ryan Shea, has introduced a simplified way to compare leading AI models by placing them on a human IQ scale. According to its latest ranking, GPT-5.5 leads the field with an IQ score of 136, making it the highest-scoring model currently listed on the platform.
Other top performers include Claude Opus 4.7 and Gemini 3.1 Pro, both at 132. They are followed by Grok 4.3 with 125 and Kimi K2.6 with 122. By translating technical benchmark performance into a single score, the platform aims to give users a more intuitive snapshot of how frontier models compare.
Built on public benchmarks instead of new testing
AI IQ does not run its own original exams. Instead, its scoring system aggregates results from 12 public benchmarks and converts them into what it calls implied IQ scores. The goal is to turn a fragmented set of technical evaluations into a format that is easier for broader audiences to understand.
The platform says it evaluates models across four dimensions: abstract reasoning, mathematical reasoning, programming, and academic reasoning. Those results are then mapped onto a human IQ bell curve, producing a single headline number. For users who do not want to parse multiple benchmark tables, this creates a more accessible comparison layer.
Tracking capability, cost efficiency, and EQ
Beyond the core ranking, AI IQ also includes additional tools such as an “IQ vs. Cost” chart, which compares model capability with cost efficiency, and a “Frontier IQ Timeline” that tracks progress over time. This expands the platform’s usefulness beyond pure performance rankings by highlighting how quickly models are improving and how much users may be paying for that intelligence.
The site also incorporates emotional intelligence data from EQ-Bench, giving users another lens beyond raw reasoning scores. While the methodology depends on existing public benchmarks and the conversion of those results into a human-style IQ metric may invite debate, AI IQ is clearly positioning itself as a more approachable dashboard for tracking competition among major AI models.
The platform is now publicly available at aiiq.org. As new models continue to launch, tools that compress complex evaluation data into a single, familiar score may become an increasingly visible way for the public to follow the pace of AI development.