Model Benchmarks

Compare the performance of different AI models across standardized benchmarks. Higher scores generally indicate better performance, but context matters for each benchmark.

Benchmarks

MedQA

Using the Medprompt method, GPT-4 achieved a score exceeding 90% on the MedQA dataset, which assesses medical question answering.

Rank	Model	Provider	Score (percent)	Date	Source
1	GPT-4	OpenAI	90.00	May 4, 2025	View
2	Mistral 7B	Mistral AI	63.00	May 4, 2025	View