Answer a few questions and get personalized model recommendations based on 540+ models.
Select one or more use cases (multi-select)
Our recommender scores models using a weighted algorithm based on your inputs. It combines real benchmark data (coding, math, reasoning scores), pricing data, speed measurements, and feature availability from Artificial Analysis and OpenRouter APIs.
The recommendations are based on objective benchmark data and pricing, not sponsored placements. However, real-world performance can vary based on your specific prompts, use patterns, and provider load. We recommend testing top picks with your actual workload.
Model data is refreshed hourly from Artificial Analysis and OpenRouter APIs, ensuring you see the latest benchmarks and pricing.
Rankings depend on your specific criteria combination. A model might excel at benchmarks but score lower if you prioritized budget or speed. Try adjusting your preferences to see how rankings change.