{"version":1,"type":"story","url":"https://digestai.news/story/researchers-release-memu-bench-to-test-gui-agents-on-long-user-session","json":"https://digestai.news/story/researchers-release-memu-bench-to-test-gui-agents-on-long-user-session.json","markdown":"https://digestai.news/story/researchers-release-memu-bench-to-test-gui-agents-on-long-user-session.md","slug":"researchers-release-memu-bench-to-test-gui-agents-on-long-user-session","headline":"Researchers release MemU-Bench to test GUI agents on long user sessions","summary":"Researchers introduced **MemU-Bench**, a new benchmark for testing GUI agents on **118 vague tasks** across **29 long-horizon user interactions**. Unlike existing tests, it simulates real-world ambiguity where users forget past steps, forcing agents to reconstruct context from noisy histories. The benchmark includes **Success Rate**, **Interaction Efficiency (IE)**, and **Success weighted by Path Length (SPL)** to distinguish agents that brute-force solutions from those that recall history accurately.\n\nThe study found that top models like **Gemini-3.1-Pro** scored only **16.4% SPL** without user history, while end-to-end GUI agents failed below **5%**. To improve performance, the team proposed **AppMem**, a framework that organizes user history into a structured memory bank. AppMem boosted **GPT-5.5’s SPL from 15.7% to 66.9%**, a **51.2% increase**, outperforming simpler summary methods by **9.3%**. The work highlights gaps in current agents’ ability to handle fragmented, real-world tasks.","keyPoints":["MemU-Bench tests GUI agents on **118 vague tasks** across **29 long user sessions** with forgotten context","Top models like **Gemini-3.1-Pro** score **16.4% SPL** without user history, end-to-end agents under **5%**","**AppMem** framework improves **GPT-5.5’s SPL by 51.2%**, from **15.7% to 66.9%**"],"whyItMatters":"MemU-Bench exposes flaws in GUI agents’ ability to handle real-world ambiguity, where users forget steps. The **AppMem** solution could guide future training to improve task recall, critical for agents in productivity tools or customer support.","category":{"slug":"agents","name":"Agents & Tools","url":"https://digestai.news/category/agents"},"entities":{"companies":[],"models":["Gemini-3.1-Pro","GPT-5.5"],"people":[]},"firstPublishedAt":"2026-11-03T05:00:00Z","updatedAt":"2026-11-03T05:00:00Z","sourceCount":2,"hasPrimarySource":true,"sources":[{"outlet":"arXiv cs.AI","title":"MemU-Bench: Benchmarking GUI Agents on Long-Horizon User Interactions","url":"https://arxiv.org/abs/2611.00191","publishedAt":"2026-11-03T05:00:00Z","type":"primary","primary":true,"lead":true},{"outlet":"arXiv cs.CL","title":"MEMENTO: Learning Context Management for Long Context Processing Agents","url":"https://arxiv.org/abs/2611.00244","publishedAt":"2026-11-03T05:00:00Z","type":"primary","primary":true,"lead":false}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"Researchers release MemU-Bench to test GUI agents on long user sessions\", 3 November 2026, https://digestai.news/story/researchers-release-memu-bench-to-test-gui-agents-on-long-user-session","publisher":"Digest AI","title":"Researchers release MemU-Bench to test GUI agents on long user sessions","datePublished":"2026-11-03T05:00:00Z","url":"https://digestai.news/story/researchers-release-memu-bench-to-test-gui-agents-on-long-user-session"},"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"}