Dark | Light
[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

[@DocAI_ocr](/creator/twitter/DocAI_ocr)
"NEWS Baidu just dropped a 0.9B model that beats GPT-4o at document parsing PaddleOCR-VL-0.9B (from Baidus PaddlePaddle team) just scored XXXXX on OmniDocBench outperforming GPT-4o Gemini-2.5 Pro and Qwen2.5-VL-72B while being XXX smaller 🤯 What blew my mind: 👉$500/mo to process 1M pages (vs $15K via APIs) 👉Fully self-hostable zero data leaves your infra 👉109 languages even Tamil Telugu Tibetan And yeah its open-source + fine-tunable for specific domains (finance legal medical.) As an agent dev THIS HIT ME. If real-world perf holds up this could replace GPT-4o Vision as the OCR backbone"  
[X Link](https://x.com/DocAI_ocr/status/1979135105546502335) [@DocAI_ocr](/creator/x/DocAI_ocr) 2025-10-17T10:39Z X followers, XXX engagements

[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

@DocAI_ocr "NEWS Baidu just dropped a 0.9B model that beats GPT-4o at document parsing PaddleOCR-VL-0.9B (from Baidus PaddlePaddle team) just scored XXXXX on OmniDocBench outperforming GPT-4o Gemini-2.5 Pro and Qwen2.5-VL-72B while being XXX smaller 🤯 What blew my mind: 👉$500/mo to process 1M pages (vs $15K via APIs) 👉Fully self-hostable zero data leaves your infra 👉109 languages even Tamil Telugu Tibetan And yeah its open-source + fine-tunable for specific domains (finance legal medical.) As an agent dev THIS HIT ME. If real-world perf holds up this could replace GPT-4o Vision as the OCR backbone"
X Link @DocAI_ocr 2025-10-17T10:39Z X followers, XXX engagements

creator/twitter::1973721648986800128/posts
/creator/twitter::1973721648986800128/posts