{"version":1,"type":"story","url":"https://digestai.news/story/alibaba-releases-industryllm-for-industrial-procurement-with-35b-param","json":"https://digestai.news/story/alibaba-releases-industryllm-for-industrial-procurement-with-35b-param.json","markdown":"https://digestai.news/story/alibaba-releases-industryllm-for-industrial-procurement-with-35b-param.md","slug":"alibaba-releases-industryllm-for-industrial-procurement-with-35b-param","headline":"Alibaba releases IndustryLLM for industrial procurement with 35B parameters","summary":"Researchers at Alibaba have introduced **IndustryLLM**, an open-weight language model designed for industrial procurement tasks. Built from **Qwen3.5-35B-A3B-Base**, the model uses **35 billion parameters** with **3 billion activated per token** and a frozen vision encoder. Its training focuses on failure-driven adaptation, combining continued pre-training (CPT) on a **100 billion-token corpus**—including **5 billion tokens** of national standards (e.g., GB/T), **10 billion tokens** of de-identified procurement records, and **60 billion tokens** of general data—alongside supervised fine-tuning (SFT) to address jargon, factual errors, and ambiguity in technical specifications.\n\nThe model achieves **2.97 percentage points** improvement in procurement-query structuring (95% CI: [2.11, 3.86]) in offline tests and **4.25% higher GMV** and **8.3% more satisfied inquiries** in live A/B experiments. Latency dropped from **6–7 seconds** to **1.5 seconds**. The weights and configurations are available on Hugging Face.","keyPoints":["IndustryLLM is a 35B-parameter model trained on 100B tokens for industrial procurement, with 3B activated per token","Offline tests show 2.97% improvement in query structuring; live A/B tests report 4.25% higher GMV and 8.3% more satisfied inquiries","Latency reduced from 6–7 seconds to 1.5 seconds, with open weights released on Hugging Face"],"whyItMatters":"IndustryLLM could streamline procurement workflows by handling technical jargon, standards, and real-world transaction data more accurately than generic models, reducing errors and speeding up decision-making in industrial settings.","category":{"slug":"research","name":"Research","url":"https://digestai.news/category/research"},"entities":{"companies":["Alibaba"],"models":["Qwen3.5-35B-A3B-Base","IndustryLLM"],"people":[]},"firstPublishedAt":"2026-09-29T04:00:00Z","updatedAt":"2026-09-29T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"sources":[{"outlet":"arXiv cs.AI","title":"IndustryLLM: Failure-Driven LLM Training for Industrial Procurement","url":"https://arxiv.org/abs/2609.31871","publishedAt":"2026-09-29T04:00:00Z","type":"primary","primary":true,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"Alibaba releases IndustryLLM for industrial procurement with 35B parameters\", 29 September 2026, https://digestai.news/story/alibaba-releases-industryllm-for-industrial-procurement-with-35b-param","publisher":"Digest AI","title":"Alibaba releases IndustryLLM for industrial procurement with 35B parameters","datePublished":"2026-09-29T04:00:00Z","url":"https://digestai.news/story/alibaba-releases-industryllm-for-industrial-procurement-with-35b-param"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}