Benchmark
Skatebench
A benchmark designed to evaluate the reasoning capability and token efficiency of AI models. It is commonly used to analyze the cost-performance ratio of different models across various reasoning tiers, enabling comparisons of how effectively models balance output quality against token consumption.
Timeline (last 90 days)
Sep 25
In the Skatebench test, from the x-high tier to the max tier, reasoning tokens per question spiked from approximately 330 to nearly 5,000 (more than 10x), while scores improved by only 1%.
Unverified50%