Last week was big for open source LLMs. We got:
- Qwen 2.5 VL (72b and 32b)
- Gemma-3 (27b)
- DeepSeek-v3-0324
And a couple weeks ago we got the new mistral-ocr model. We updated our OCR benchmark to include the new models.
We evaluated 1,000 documents for JSON extraction accuracy. Major takeaways:
- Qwen 2.5 VL (72b and 32b) are by far the most impressive. Both landed right around 75% accuracy (equivalent to GPT-4o’s performance). Qwen 72b was only 0.4% above 32b. Within the margin of error.
- Both Qwen models passed mistral-ocr (72.2%), which is specifically trained for OCR.
- Gemma-3 (27B) only scored 42.9%. Particularly surprising given that it's architecture is based on Gemini 2.0 which still tops the accuracy chart.
The data set and benchmark runner is fully open source. You can check out the code and reproduction steps here:
- https://getomni.ai/blog/benchmarking-open-source-models-for-...
- https://github.com/getomni-ai/benchmark
- https://huggingface.co/datasets/getomni-ai/ocr-benchmark
Comments URL: https://news.ycombinator.com/item?id=43549072
Points: 61
# Comments: 13
Chcete-li přidat komentář, přihlaste se
Ostatní příspěvky v této skupině

Article URL: https://thenewstack.io/nvidia-finally-adds-native-python-support-to-cuda/
Comments URL:
For popular threads with a lot of comments, collapsing a top comment leads to high CPU usage and can cause browsers (I tried in Chrome and Vivaldi) to freeze up.
For example, this very popular t

Article URL: https://lyra.horse/css-clicker/
Comments URL: https://news.ycombinator.com

