Le Chonk cooked on visual grounding.
Pinpoints exactly where things are in an image.
42% on Dense 200, ahead of even frontier closed models..
Mistral's Le Chonk scores 42% on Dense 200 visual grounding benchmark
AISummary
Mistral AI says its Le Chonk model pinpoints exactly where objects are in an image. It scores 42% on the Dense 200 benchmark, which the post says is ahead of even frontier closed models.
Post on XView on X
Source: Mistral AI · x.comPublished · added here