Model Advisor
Pick the right AI model for the job
Paste a prompt or describe your use case. We rank every model in the live catalog on quality, cost and speed, and tell you why.
Runs locally in your browser. No AI call, no cost, nothing sent anywhere.
What matters most?
* The catalog has no measured latency data. The speed score is an estimate based on model tier conventions: small tiers reply faster, reasoning models slower.
Add some prompt text to see recommendations.
How the ranking works
- Every model whose context window cannot hold your input plus its reply is disqualified, and named for you.
- Cost is the real estimated price of one call at your token counts, not a headline per-million rate.
- Quality blends the catalog capability tags with model tier and price, and tags always outweigh inferences.
- Speed is inferred from model tier because the catalog carries no latency measurements. It is labelled everywhere it appears.
- Nothing here is sponsored. Rankings use catalog data only, never affiliate relationships.