zai-org/GLM-OCR
Primitive: /extract · Extract ·
GLM-OCR
👋 Join our WeChat and Discord community
MultimodalMultilingualLong contextEntities
Overview
Hardware: — drives latency, throughput & cost
| Size | 1.3B params |
|---|---|
| Tasks | /extract |
| License | mit |
| Languages | zh, en, fr, es, ru, de, ja, ko |
| Latency | 16.8 s |
| Throughput | 662 tok/s |
| Cost | $0.336 /1M tok |
Cost is approximate — computed from list GPU prices; your actual price depends on the provider you deploy SIE with.
Extraction
| Output kinds | Entities |
|---|---|
| Inputs | image |
| Max sequence length | 16,384 |
Benchmarks
olmOCR-Bench
Document text extraction accuracy across arxiv math, old scans, multi-column layouts, and tables
Corpus: 1,403 Queries: 1,403
default
Quality
accuracy 0.7380
Performance L4 b1 c4
Extract 0.4 mpix/s
Extract p50 16.8s
default_limit-50
Quality
accuracy 0.6943
Compare (0)Compare →