New OCR models expand document-processing options
A post published on September 18, 2026 presents WeVisDoc 2B and 4B under Apache 2.0, and jina-ocr-v1 under a non-commercial license. It reports a 95.38 score for WeVisDoc-4B on OmniDocBench v1.6 and identifies NaviDC-OCR as the current SOTA.
Source: WeVisDoc e jina-ocr-v1 são novos modelos de OCR (x.com). Text prepared with AI from this source.
What happened and what to do
On September 18, 2026, a post presented Tencent's WeVisDoc 2B and 4B under the Apache 2.0 license, and jinaai's jina-ocr-v1 under a non-commercial license. According to the post, WeVisDoc-4B scored 95.38 on OmniDocBench v1.6, while NaviDC-OCR is the current SOTA. The model options and license terms can inform an OCR comparison.
A company can evaluate these options using documents representative of its operations, checking extraction quality, licensing requirements, and infrastructure compatibility. A response could include an OCR pipeline with validation, monitoring, and routing for low-confidence cases, without assuming that a benchmark result will carry over to the company's own documents.
How the consultancy can help
Wendelmaques can assess the document workflow, licensing requirements, and infrastructure, define an evaluation scope, and implement and operate an OCR solution with a pipeline, validation, and monitoring.
Next step
Send a short description of the documents and process you want to improve to receive a scoped proposal.
Let's assess your project
I reply with scope and a proposal.