Comment by refulgentis

6 days ago

OCR is off the charts good on every metric you can think of.

LLMs are a near-afterthought at this point if you don’t have data residency requirements. I love them and they’re slightly underrated, their models are consistently well-trained, open, but as you note, behind. There is no metric that will say they’re ahead in anything.

This. Best OCR provider by measure and it’s been for years

  • Hmm, not sure I'd agree. I really like google's offering there (they suck at coding agents but their OCR is good value for money - well up till the latest flash model which has got wicked expensive). See also https://www.ocrarena.ai/leaderboard I know these leaderboards are iffy, but at least my experience has been somewhat similar.

    • Thank you for sharing this, I’ve had some disquiet around the release blogs…something felt cherry-picked and I didn’t know there was a 3rd party source for evals, that settles it easily - like you said they can be iffy but there’s a clear enough gap and large set of models for me to let go of the idea it’s the best by some margin.