Title: GLM-OCR vs GLM-4.7-Flash: Parametry, Licence %E2%80%94 LLM EU
Canonical: https://llmeu.com/cs/compare/glm-ocr-vs-glm-4-7-flash/
Markdown mirror: https://llmeu.com/cs/compare/glm-ocr-vs-glm-4-7-flash.md
Machine-readable: https://llmeu.com/llms.txt (index) - https://llmeu.com/llms-full.txt (full catalogue) - https://llmeu.com/llms-catalogue.json

# GLM-OCR vs GLM-4.7-Flash

Porovnejte dva a%C5%BE %C4%8Dty%C5%99i modely vedle sebe podle pol%C3%AD, kter%C3%A1 rozhoduj%C3%AD o nasazen%C3%AD: kontext, cena, region, licence, jazyky, hosting a co je ov%C4%9B%C5%99eno.

Tam, kde se dva modely li%C5%A1%C3%AD, tabulka uk%C3%A1%C5%BEe rozd%C3%ADl a skon%C4%8D%C3%AD. LLM EU nezve%C5%99ej%C5%88uje odznak v%C3%ADt%C4%9Bze bez dolo%C5%BEen%C3%A9ho sk%C3%B3re pro danou %C3%BAlohu a chyb%C4%9Bj%C3%ADc%C3%AD %C3%BAdaj z%C5%AFst%C3%A1v%C3%A1 chyb%C4%9Bj%C3%ADc%C3%AD, m%C3%ADsto aby se stal nulou.

## Kdy pou%C5%BE%C3%ADt

### [GLM-OCR](/cs/models/glm-ocr/)

#### Siln%C3%A9 str%C3%A1nky

- purpose-built document OCR rather than general vision
- runs on modest hardware and through common inference servers
- MIT licence

#### Omezen%C3%AD

- text extraction only; it is not a general chat or reasoning model
- vendor parameter count (0.9B) differs from the published safetensors total (1.33B)
- no OCR benchmark has been run by LLM EU

#### %C3%9Alohy

[Zpracov%C3%A1n%C3%AD obrazu](/cs/tasks/vision/), [Vyhled%C3%A1v%C3%A1n%C3%AD (RAG)](/cs/tasks/rag/), [Pr%C3%A1vn%C3%AD n%C3%A1vrhy](/cs/tasks/legal-draft/), [N%C3%ADzk%C3%A9 n%C3%A1klady](/cs/tasks/cheap/), [Shrnut%C3%AD](/cs/tasks/summarization/)

### [GLM-4.7-Flash](/cs/models/glm-4-7-flash/)

#### Siln%C3%A9 str%C3%A1nky

- MIT licence with a permissive commercial position
- high published maths score for a 30B-class model
- roughly 200K-token context window

#### Omezen%C3%AD

- benchmark table is self-reported by the vendor
- English and Chinese are the only declared languages
- no evaluation has been run by LLM EU

#### %C3%9Alohy

[Uva%C5%BEov%C3%A1n%C3%AD](/cs/tasks/reasoning/), [K%C3%B3dov%C3%A1n%C3%AD](/cs/tasks/coding/), [N%C3%ADzk%C3%A9 n%C3%A1klady](/cs/tasks/cheap/), [Shrnut%C3%AD](/cs/tasks/summarization/)

**%C3%9Alohy:** [N%C3%ADzk%C3%A9 n%C3%A1klady](/cs/tasks/cheap/), [Shrnut%C3%AD](/cs/tasks/summarization/)

## Porovnat modely

- [Bloom 1b1 vs GLM-OCR](/cs/compare/bloom-1b1-vs-glm-ocr/)
- [Command R (08-2024) vs GLM-4.7-Flash](/cs/compare/command-r-08-2024-vs-glm-4-7-flash/)
- [NVIDIA Cosmos Reason2 2B vs GLM-OCR](/cs/compare/cosmos-reason2-2b-vs-glm-ocr/)
- [DeepSeek Coder V2 Lite Instruct vs GLM-4.7-Flash](/cs/compare/deepseek-coder-v2-lite-instruct-vs-glm-4-7-flash/)
- [DeepSeek R1 Distill Qwen 1.5B vs GLM-OCR](/cs/compare/deepseek-r1-distill-qwen-1-5b-vs-glm-ocr/)
- [DeepSeek-R1-Distill-Qwen-32B vs GLM-4.7-Flash](/cs/compare/deepseek-r1-distill-qwen-32b-vs-glm-4-7-flash/)