Model
Gemini 3.1
A version of the Gemini series of large language models released by Google DeepMind. As the predecessor of Gemini 4, it scored 71% on the knowledge work benchmark.
Timeline (last 90 days)
Sep 23
Under the rebuilt R configuration, the sign between Gemini 3.1 and Gemini 3.7 was reversed, and Persistence failed.
Unverified50%