Published on February 3, 2026, the announcement describes a 0.9B multimodal OCR model using the GLM-V architecture and an MIT license. The author says it ranks first on OmniDocBench v1.5, with a score of 94.62.
The announcement, published on February 3, 2026, presents GLM-OCR as a 0.9B multimodal OCR model based on the GLM-V architecture and released under the MIT license. The author says the model ranks first on OmniDocBench v1.5, with a score of 94.62.
Engineers assessing OCR can consult the model page and compare the claim with the original benchmark's results and criteria. If using AI to study or apply the material, avoid submitting confidential documents without authorization and use appropriate controls for organizational data.