GLM-OCR
No written verdict yet. Here is what commenters said.
6recommend
1warn against
Feb 11, 2026first mention
Aug 24, 2026latest mention
Why people recommend it
“Another decent option is GLM OCR”
“you can switch between DeepSeek OCR 2, GLM-OCR, dots.mocr, Paddle OCR VL, PP-OCRv6, etc.”
Show 3 more
“I've had good results with glm-ocr on complex documents”
“The one I have been recommending lately is GLM-OCR”
“GLM-OCR is the best single model I have found for OCR over documents with good table extraction/formatting.”
Caveats and dissent
“GLM-OCR didn't work because it returns all text on the screen”
Often compared with
Who recommends it in their own words
Threads where it came up
- Gave Claude photographic memory for $0.0002/screenshot 3 mentions
- OCR It – pull text out of un-copyable documents for your LLM 1 mention
- GPT 5.6 Sol is the best "vision" model OpenAI ever released 1 mention
- Launch HN: Parsewise (YC P25) – Reason Across Documents with an API 1 mention
- Show HN: I Built Paul Graham's Intellectual Captcha Idea 1 mention
- Google releases Gemma 4 open models 1 mention
Also written as glm-ocr. Data: glm-ocr.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 11
“Another decent option is GLM OCR”
“you can switch between DeepSeek OCR 2, GLM-OCR, dots.mocr, Paddle OCR VL, PP-OCRv6, etc.”
“I've had good results with glm-ocr on complex documents”
“The one I have been recommending lately is GLM-OCR”
“GLM-OCR is the best single model I have found for OCR over documents with good table extraction/formatting.”
“Try just GLM-OCR if you want to get started quickly”
Update: I switched to Qwen3 VL 2B (`qwen3-vl-2b-instruct-mlx@bf16`) which is 2.5× faster than 4B (11s vs 18s per screenshot) and my meager M1 Pro is able to keep up without the fans spinning 100% of the time.
“GLM-OCR didn't work because it returns all text on the screen”
“GLM-OCR seems like it would excel at this”
“I've been using it for digitizing various hand-written notes”
> I burn up the entire context window for 30 pages of text We analyze 200 page contracts no problem I think you're doing it wrong or in an antiquated way (until context window sizes improve) Are you doing this programmatically at all or are you doing something closer to dropping a contract into a chat window? We use a main agent to classify the pages and we build subagents that are familiar with page classifications