AIMedmarks adds medical benchmark results for mid-size open-source AI models
SophontAI says Gemma 4 31B leads in this size class. The new Medmarks results cover Gemma, Qwen, Muse and Nemotron.
TLDR
SophontAI announced additional Medmarks results on September 21, 2026, covering mid-size open-source language models. It describes Medmarks as an open benchmark suite and leaderboard for medical capabilities, and says Gemma 4 31B leads in that size class. A researcher working on Medmarks also said evaluations of Astra and Fable were underway for the next release.
Medmarks adds medical benchmark results for mid-size open-source AI models
SophontAI says Gemma 4 31B leads in this size class. The new Medmarks results cover Gemma, Qwen, Muse and Nemotron.
