AI Briefing#ibm-research
The latest AI and developer news about #ibm-research, with the original source and a short summary.
The latest AI and developer news about #ibm-research, with the original source and a short summary.
A fine-tuned 20B model using a designer-coder split and check-repair harness achieves a 59.7 mean score on PresentBench, ranking second overall and first in Economics and Education.
IBM's open-source CUGA agent harness achieved an 81.1% task completion rate on the AppWorld benchmark, significantly outperforming Hermes and Deep Agents using the same GLM-5.3-flash model.
The method uses GPT-4.1 to extract behavioral guidelines from agent trajectories, improving reliability and accuracy without weight updates or human annotation.
IBM Research defined the 'Consistency Gap' metric to measure the consistency of repeated AI agent executions and released a tool to improve it.
IBM and UC Berkeley have released MAST, a methodology that systematically analyzes failure patterns in IT automation agents.
AssetOpsBench, an AI agent benchmark for evaluating complex operational environments in industrial settings, has been released.
CUGA, an open-source AI agent framework that performs complex web and API tasks, has been released.
IBM Research's lightweight time series model PatchTSMixer has been integrated into the HuggingFace Transformers library.