Follow us to discover more Baidu tech insights.
IDC’s April 2024 report shows China’s automotive cloud market reaching 6.51 billion RMB in the second half of 2024, with a 27.4% YoY growth, and highlights Baidu Cloud’s 34.5% share in the 16.34 billion RMB autonomous‑driving solution market, driven by AI advances and expanding compute investments.
This article details Baidu's log platform point‑governance practice, explaining why uncontrolled event logging inflates storage and compute costs, and describing a three‑stage solution—manual, semi‑automatic platform, and full‑lifecycle standardization—that uses anomaly detection, automated workflows, and IM bots to achieve massive PV reduction and annual cost savings.
Baidu’s SRE digital‑immune system, evolved into an AI‑powered intelligent immunity platform, quantifies and mitigates risk across thousands of services by integrating data‑driven monitoring, rule‑based detection, and large‑model GraphRAG knowledge mining, cutting degradation cases by ~40% and shifting operations from reactive troubleshooting to proactive, data‑centric quality assurance.
At the AIIA’s 14th plenary meeting in Nanjing, the FactTesting benchmark released its Q1 2025 results, evaluating over 200 large models and highlighting Baidu’s Wenxin 4.5 and Wenxin X1 as leaders in basic and reasoning capabilities, while outlining the expanded multimodal and agent testing roadmap for the year.
PaddlePaddle Framework 3.0 delivers five breakthroughs—dynamic‑static unified automatic parallelism, integrated training‑inference pipelines, high‑order scientific differentiation, a neural‑network compiler with automatic operator fusion, and streamlined heterogeneous chip adaptation—drastically reducing development effort, boosting training speed, and expanding compatibility for large‑scale AI models.
On April 2, Baidu released its Wenxin X1 large model on the Qianfan platform, offering enterprise users and developers a multimodal, deep‑thinking AI with superior math, coding, and reasoning scores, low token‑price API access, batch inference, one‑click distillation, and rapid RAG/Agent application building.
COBRA, Baidu’s new generative retrieval framework, unifies sparse ID generation and dense vector encoding through a cascaded architecture that first predicts hierarchical IDs then refines them into dense representations, achieving state‑of‑the‑art recall, NDCG and conversion gains across public benchmarks and large‑scale advertising production.
DeepSeek‑VL2 is a state‑of‑the‑art multimodal model built on a Mixture‑of‑Experts architecture that combines a SigLIP‑L vision encoder with dynamic tiling, a two‑layer VL adaptor, and a DeepSeek‑MoE language model using Multi‑head Latent Attention, trained in three stages on diverse visual‑language and text data, and achieving strong results on benchmarks such as DocVQA and TextVQA, with full implementation and inference code available in PaddleMIX.
The article provides a detailed technical overview of the Turing Data Finder (TDF) platform, describing its background, core components, data schema, ingestion workflow, and a suite of growth‑analysis features such as event, retention, funnel, path, component, distribution, and attribution analysis, while also outlining performance‑optimisation techniques and future development directions.
Baidu officially launched the Wenxin 4.5 and X1 large language models, showcasing native multimodal foundations, advanced attention masks, heterogeneous expert extensions, and tool‑calling capabilities, while offering low‑cost API access on the Qianfan platform and outlining the underlying technical innovations that drive their performance gains.
