Introducing UsageBench on BridgeBench.
If an AI lab ever rug pulls your subscription, you will see it here.
Day one is 100%. Then we keep retesting the same plan at the same price, with 5 hour and weekly limits tracked separately.
Below 100%, you are getting less. Above 100%, you are getting more.
Starting with Claude Max 20x and ChatGPT Pro. OpenAI changes Pro usage tomorrow, and we already have the baseline.
Which plan should we track next?