Did they make significant improvements in OCR 3? The quality I was getting from Mistral OCR 2 was nowhere near as good as what I could get from just sending the same files to Claude Sonnet via an API call.

I have been finding Voxtral useful though.

Better than Qwen? I guess the best overall is Gemini, right?

probably yes. considering that even some of their non-ocr models can recognize my shitty handwritten math

[dead]