Comment on From GPT-4 to GPT-5: Measuring progress through MedHELM [pdf]parentComments−fertrevinoOP1yThat would be an interesting extension. MedGemma isn't part of the original benchmark either [1]. Since Gemini 2.0 Flash is on 6th place, expectations are for MedGemma to achieve higher than that :)[1]https://crfm.stanford.edu/helm/medhelm/latest/#/leaderboard
Comments
That would be an interesting extension. MedGemma isn't part of the original benchmark either [1]. Since Gemini 2.0 Flash is on 6th place, expectations are for MedGemma to achieve higher than that :)
[1]https://crfm.stanford.edu/helm/medhelm/latest/#/leaderboard