TechniqueRAG1 recent entries13 Apr 2026IBM demonstrates extreme scale with a 100B vector databaseIBM Research demonstrated the capability to store and search across a database of 100 billion vectors, showcasing extreme-scale vector storage for AI applications. This work addresses the growing need
TechniqueAgents1 recent entries25 Aug 2026Granite 4.2 brings native reasoning to enterprise agentsIBM has released Granite 4.2, an open‑source agentic AI model that adds native reasoning, tool use, coding instruction following and speech capabilities to enterprise agents. The suite of pre‑built mo
TechniqueFine-tuning1 recent entries15 Apr 2026Mid-training is essential for LLM reasoning, IBM study showsIBM Research has demonstrated that mid-training — a dedicated training phase between initial pre-training and fine-tuning — is essential for improving reasoning capabilities in large language models (
TechniqueMultimodal2 recent entries12 Aug 2026DocLang: a markup language for LLMsIBM has introduced **DocLang**, a new markup language specifically designed for large‑language models (LLMs). Developed by lead researcher Peter Staar and IBM’s document‑parsing team, DocLang aims to →25 Aug 2026Granite 4.2 brings native reasoning to enterprise agentsIBM has released Granite 4.2, an open‑source agentic AI model that adds native reasoning, tool use, coding instruction following and speech capabilities to enterprise agents. The suite of pre‑built mo