Comment by armarr

12 hours ago

GLM was twice as verbose running the Artificial Analysis benchmark. So it ends up being more expensive