Baidu just open-sourced Unlimited OCR 2, a 3B document-parsing model that can ingest thousands of pages in a single request. By revamping attention with SR-SWA, it keeps memory and latency stable over massive documents, automatically stitching together cross-page tables and split text. It easily fits on consumer GPUs, meaning you no longer have to chop books into pieces just to digitize them.
Baidu just introduced Unlimited OCR, a 3B MoE document parser that stops memory bloat across 40+ pages. By using an R-SWA mechanism to cap the text KV cache to the last 128 tokens, the model processes long documents at a completely constant speed.
Baidu just open-sourced NAVA, a 6.3B-parameter model that jointly generates 720p video and stereo audio from a single prompt instead of relying on post-production dubbing. It supports reference-audio voice control and reaches state-of-the-art lip-sync benchmarks with up to 5x fewer parameters than rival open baselines.
Baidu Cloud just launched DuClaw, a turnkey OpenClaw service that runs straight in the browser. Users can run full autonomous agent workflows without provisioning servers or API keys, complete with built-in Baidu search skills and instant switching between models like DeepSeek and Kimi-K2.5.
Baidu just open-sourced its ERNIE 4.5 model collection on Hugging Face, spanning lightweight text models to massive multimodal models reaching 424B parameters. The release provides checkpoints in both PyTorch and PaddlePaddle alongside dedicated Thinking variants, giving open-source developers a serious heavyweight alternative.
Baidu just launched an e-commerce livestream featuring top host Luo Yonghao's digital avatar, powered by its ERNIE foundation model. It was the first livestream to feature two digital avatars co-hosting and interacting in real time. The 6+ hour stream pulled in over 13 million views and generated more than 55 million RMB in GMV.
ERNIE 4.5 just swept GPT-4.5 in a three-game chess showdown, winning all three matches decisively. The model demonstrated tighter instruction-following and sharper positional decision-making, even sealing the final match with a 10-move checkmate. It is a reminder that frontier logic tests do not always belong exclusively to OpenAI.
Baidu just unveiled its ChatGPT rival Ernie Bot in Beijing, but reliance on pre-recorded video demos instead of a live test triggered an investor selloff. Shares dropped 6.4% in Hong Kong, wiping out $3 billion. Why it matters: China's biggest AI launch stumbled right out of the gate, with CEO Robin Li admitting the tool is not yet ready.
