24 comments

  • molf2 hours ago
    What is needed to evaluate OCR for most business applications (above everything else) is accuracy.<p>Some results look plausible but are just plain wrong. That is worse than useless.<p>Example: the &quot;Table&quot; sample document contains chemical substances and their properties. How many numbers did the LLM output and associate correctly? That is all that matters. There is no &quot;preference&quot; aspect that is relevant until the data is correct. Nicely formatted incorrect data is still incorrect.<p>I reviewed the output from Qwen3-VL-8B on this document. It mixes up the rows, resulting in many values associated with the wrong substance. I presume using its output for any real purpose would be incredibly dangerous. This model should not be used for such a purpose. There is no winning aspect to it. Does another model produce worse results? Then both models should be avoided at all costs.<p>Are there models available that are accurate enough for this purpose? I don&#x27;t know. It is very time consuming to evaluate. This particular table seems pretty legible. A real production grade OCR solution should probably need a 100% score on this example before it can be adopted. The output of such a table is not something humans are good at reviewing. It is difficult to spot errors. It either needs to be entirely correct, or the OCR has failed completely.<p>I am confident we&#x27;ll reach a point where a mix of traditional OCR and LLM models can produce correct and usable output. I would welcome a benchmark where (objective) correctness is rated separately from of the (subjective) output structure.<p>Edit: Just checked a few other models for errors on this example.<p>* GPT 5.1 is confused by the column labelled &quot;C4&quot; and mismatches the last 4 columns entirely. And almost all of the numbers in the last column are wrong.<p>* olmOCR 2 omits the single value in column &quot;C4&quot; from the table.<p>* Gemini 3 produces &quot;1.001E-04&quot; instead of &quot;1.001E-11&quot; as viscosity at T_max for Argon. Off by 7 orders of magnitude! There is zero ambiguity in the original table. On the second try it got it right. Which is interesting! I want to see this in a benchmark!<p>There might be more errors! I don&#x27;t know, I&#x27;d like to see them!
  • ArcaneMoose13 hours ago
    I&#x27;ve been really impressed with this model specifically because of how insanely cheap it is: <a href="https:&#x2F;&#x2F;replicate.com&#x2F;ibm-granite&#x2F;granite-vision-3.3-2b">https:&#x2F;&#x2F;replicate.com&#x2F;ibm-granite&#x2F;granite-vision-3.3-2b</a><p>I didn&#x27;t expect IBM to be making relevant AI models but this thing is priced at $1 per 4,000,000 <i>output</i> tokens... I&#x27;m using it to transcribe handwritten input text and it works very well and super fast.
    • rubikscubeguy8 hours ago
      I&#x27;m the dev who made this:) We are looking into adding granite!
    • nicman237 hours ago
      English only :( . it seems only 2 orders of magnitude larger models have support for ie greek :&#x2F;
    • irjustin10 hours ago
      Thanks for this! Will test this model out because we do a lot of in between steps to get around the output token limits.<p>Super nice if it worked for our use case to simply get full output.
  • poulpy1231 hour ago
    I&#x27;m very impressed by the models, to the point I was wondering if they were really converting the pdf or just reading the content. I tried on documents in french, english and spanish, very heaving on graphics and with complex layouts (boardgame, flyer, book about rust), and I wasn&#x27;t expecting anything great. Especially some models were showing symbols and smileys quite close from the original.<p>I noticed that some models were resisting better to faking data than other, especially I saw that in a sentence cut from the document, GPT5 was inventing the end of the sentence and opus was properly showing it cut.<p>I didn&#x27;t try with my writing but in the playground there is one example and some models read it better than me.<p>I wish the output would show the confidence of the model on each part. I think it would help immensely.<p>Note that sometimes a model get stuck in a loop, preventing to vote and to see which model is which
  • deaux9 hours ago
    I suggest you make explicit the assumption that this website is specifically about <i>English</i> text. Otherwise the leaderboard is pretty meaningless, with extreme differences in performance across other scripts - and potentially even languages such as Vietnamese or Czech which use Latin but have lots of accents.
    • rubikscubeguy7 hours ago
      Hey! I&#x27;m the dev who made this:) I think that you are right, data will bias towards english because we have a dataset that people can use that is in english. But you can also upload non-english docs into the battle mode as well as the playground!
      • skissane7 hours ago
        LMArena splits their leaderboard by language: maybe you should consider doing the same thing<p>I assume to do that you’d need another model to do language detection on the inputs and&#x2F;or outputs; but a language detection model can be a lot cheaper than an OCR model or an LLM
    • hdjrudni8 hours ago
      That&#x27;s unfortunate because I have a bunch of photos with handwritten German on the back that I need to transcribe, and seeing as that I can&#x27;t read German I can&#x27;t really do it by myself either.
      • maverwa1 hour ago
        from my first tests it does fine with german, at least for the gastly &quot;handwritten&quot; font the restaurant menu I used for the test uses.
      • deaux7 hours ago
        I reckon performance on German will be similar to English, the only real difference is the umlauts and those are very consistent. Not sure how it will do on the ß.
      • nicman237 hours ago
        qwen 3.5 vl instruct on openrouter is damn cheap - and works quite well with non english stuff.<p>i have it verify some stamps which are quite messy and sometimes obscured and honestly some i could not even read.
  • est7 hours ago
    Offtopic, but what&#x27;s the best OCR that can run offline on browsers with js&#x2F;wasm with reasonable CPU&#x2F;memory cost?<p>Working on a hobby project that interacts with user handwriting on &lt;canvas&gt;. Tried some CNN models for digits but had trouble with characters.
    • yorwba5 hours ago
      If the text is written interactively on the canvas (as opposed to extraction from pixels) this task is known as &quot;online handwriting recognition&quot; (&quot;online&quot; because you can watch the text being formed incrementally, which makes it easier to e.g. distinguish individual strokes.)<p>I don&#x27;t know what the state of the art is, but an old model for digitizer pens might not do so bad either.
  • cdrini11 hours ago
    There have been such a large number of OCR tools pop up over the past ~year; sorely in need for some benchmarks to compare them. Would love to see support for normal OCR tools like tesseract, EasyOCR, Microsoft Azure, etc. I&#x27;m using these for some projects, and my experiments with VLMs for OCR have resulted in too much hallucination for me to switch. Benchmarks comparing across this aisle would be incredibly useful.
  • daemonologist11 hours ago
    A limitation of this leaderboard approach that I want to point out is that while the large general-purpose LLMs can make greater leaps of inference (on handwriting and poor quality scans), and almost always produce better layouts and more coherent output, they can also sometimes be <i>less correct</i>. My experience is that they&#x27;re more prone to skipping or transposing sections of text, or even hallucinating completely incorrect output, than the purpose-trained models. (A similar comparison can be made in turn to the character- or word-based OCR approaches like Tesseract, which are even less &quot;intelligent&quot; but also even less prone to those malbehaviors.)<p>Also, some of the models are prone to infinite loops and I suspect this is not being punished appropriately; the frontend seems to get into a bad state after around 50k characters, which prevents the user from selecting a winner. Probably would be beneficial to make sure every model has an output length limit.<p>Still, a really cool resource - I&#x27;m looking forward to more models being added.
    • rubikscubeguy11 hours ago
      Totally agree w&#x2F; your first point! For the looping, we just added a stop condition for now in battle mode, and you can still vote on the other model afterwards. A bit of a hard problem to solve. We will add more models!
  • tarruda4 hours ago
    Interesting that the 8B of the Qwen3-VL family 9th place, above a few proprietary models. This thing can run locally with llama.cpp on modest hardware.
  • zzleeper14 hours ago
    Love this! Would have liked to see something like textract for a pre-LLM benchmark (but of course that&#x27;s expensive), and also a distinction between handwritten text and printed one.<p>But still, this is incredibly useful!
  • ComputerGuru9 hours ago
    Two suggestions:<p>UX on mobile isn’t great. It wasn’t obvious to me where the second model output was and I was thrown off even more so because the option to vote for model 1 output was presented without ever even seeing model two output.<p>Second suggestion would be to install a MathJax plugin so one can properly rate mathematical equations and formulas. Raw LATeX is easy to mistake and it makes comparing between LATeX and Unicode outputs hard.
    • rubikscubeguy7 hours ago
      Hey! Dev who made this here. I hear you on the mobile UX, it&#x27;s on my docket of things to fix. Same with math plugin! Thanks for the suggestions.
  • densekernel5 hours ago
    Any plans to add Document Pre-trained transformer-2 (DPT-2) from <a href="https:&#x2F;&#x2F;landing.ai&#x2F;" rel="nofollow">https:&#x2F;&#x2F;landing.ai&#x2F;</a>?
  • hakunin10 hours ago
    Would be great to compare these against Apple’s LiveText. This project now supports it: <a href="https:&#x2F;&#x2F;github.com&#x2F;mkyt&#x2F;OCRmyPDF-AppleOCR" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;mkyt&#x2F;OCRmyPDF-AppleOCR</a><p>I’ve had great results locally. Albeit you need macOS &gt;=13 for this.
  • ajmurmann8 hours ago
    Really like the idea. Unfortunately, my first upload is still spinning on one of the models about 5 minutes in. Clicking &quot;Stop Battle&quot; seems to do nothing either
    • rubikscubeguy8 hours ago
      Hey, I&#x27;m the dev who built this! Looking into it. Wondering if it&#x27;s because of load due to this post.
  • wener12 hours ago
    Really hope there is a layout mode or ocr with bbox mode, I want to see the model restore the whole page.
    • rubikscubeguy11 hours ago
      yeah, that would be a cool long term goal
  • coulix8 hours ago
    We need to see Landing.ai DPT-2, from my tests its the best in term of ability to extract structure from complex tables so far.
  • prodigycorp9 hours ago
    This needs a &quot;both are bad&quot; button. There are some generations where I cannot rightfully beats the other.
  • codeddesign13 hours ago
    Most of these are general LLM’s and not specifically OCR models. Where is Google Vision, Mistral, Paddle, Nanonets, or Chandra??
    • kbyatnal12 hours ago
      We wanted to keep the focus on (1) foundation VLMs and (2) open source OCR models.<p>We had Mistral previously but had to remove it because their hosted API for OCR was super unstable and returned a lot of garbage results unfortunately.<p>Paddle, Nanonets, and Chandra being added shortly!
    • rubikscubeguy12 hours ago
      nanonets is live now!
  • fzysingularity14 hours ago
    FYI one of the models on the battle was pretty slow to load. Are these also being rated on latency or just quality?
    • kbyatnal12 hours ago
      Ultimately, there’s some intersection of accuracy x cost x speed that’s ideal, which can be different per use case. We’ll surface all of those metrics shortly so that you can pick the best model for the job along those axes.
    • andrewlu012 hours ago
      ideally we want people to rate based on quality - but i imagine some of the results are biased rn based on loading time
      • hdjrudni8 hours ago
        That&#x27;s an easy fix if you wait for the slowest one and pop them both in at the same time, no?
  • mkolodny9 hours ago
    This is super helpful :) Curious about Grok as well!
    • rubikscubeguy7 hours ago
      Hello! Dev who made this here. Working on adding grok.
  • krashidov14 hours ago
    I would be curious to see how Sonnet does. Their models are pretty solid when it comes to PDFs
    • kbyatnal12 hours ago
      Sonnet&#x2F;Opus is being added shortly!
    • rubikscubeguy12 hours ago
      sonnet and opus are live now :)
  • ianhawes14 hours ago
    Please add Chandra by Datalab
  • aixpert6 hours ago
    Opus is multimodal??
  • arathis14 hours ago
    Claude would be good!
    • kbyatnal12 hours ago
      Claude coming shortly (in the next ~1 hour)
    • rubikscubeguy12 hours ago
      claude is live now!
  • dang15 hours ago
    [under-the-rug stub]<p>[see <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=45988611">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=45988611</a> for explanation]
    • ylhert3 days ago
      We&#x27;ve got like 10 LLM arenas but nothing for OCR yet, really hope this takes off!
    • athoscouto3 days ago
      Nice! Would love to see Azure Document Intelligence on this
    • profburial3 days ago
      This is a killer idea!