Question 1

How does the model picker choose models?

Accepted Answer

An AI model reads your task description and picks the best-fit models from TokenCost's live catalog of 140+ models, weighing your stated priority (cost, quality, or balance), context window fit, and current per-token prices. Cost estimates are computed from your token and call inputs against listed prices.

Question 2

Are the monthly cost estimates accurate?

Accepted Answer

Estimates use your average input/output tokens per call and calls per day against current listed API prices. They exclude prompt caching discounts, batch discounts, and reasoning-token overhead, so treat them as a baseline; real bills can be lower with caching or higher with reasoning models.

Question 3

Is my task description stored?

Accepted Answer

Your description is sent to an AI model to produce the recommendation and is not published or used to build profiles. Avoid pasting secrets or confidential data.

Which model should you use?

Related Tools