#baidu

ERNIE 5.1 API: Baidu's Models for Production ERNIE 5.1 is Baidu's current flagship text...

Beijing's approval of commercial robotaxi fares for Baidu's Apollo Go and Pony.ai in August 2026 is...

Unlimited OCR is Baidu's open-source end-to-end OCR model that parses dozens of pages in a single forward pass. This 2026 guide explains how Reference Sliding Window Attention (R-SWA) works, why it tops OmniDocBench, and how to run unlimited ocr locally with Transformers or SGLang.

Unlimited OCR replaces the decoder's attention layers with a design that holds the key-value cache at a constant size, so transcription speed does not degrade as the document gets longer.

中文版: 中文版本 TL;DR. Mainland Chinese consumer attention concentrates on three platforms with almost no...

Baidu open-sourced Unlimited-OCR on June 22, 2026 (MIT) — a 3B mixture-of-experts model with 500M active params that parses 40+ page documents in a single forward pass. Its new Reference Sliding Windo

Originally published at norvik.tech Introduction An in-depth analysis of Kunlunxin's...

Baidu Unlimited OCR: How Reference Sliding Window Attention Rewrites Document AI Baidu...

Baidu just released Unlimited OCR — a groundbreaking MIT-licensed model that reads dozens of pages in a single pass. But there's a catch: you need a Chinese phone number to use Baidu's API. Here's how to bypass that.

Baidu's Unlimited-OCR (MIT license) is exciting — but international developers can't easily register. Here's the workaround.

Baidu's ERNIE 4.0 is one of China's most widely deployed AI models. ...

GoldBean 封装了百度 AI 13 个核心 API（OCR + 翻译 + NLP），无需中国手机号、无需预存款，$0.01/次起，全球可用。Python/Node.js/cURL 示例齐备。