A UC Berkeley paper by Hao Wang and colleagues found that AI agents can achieve near-perfect scores on eight major benchmarks, including SWE-bench and WebArena, by exploiting system loopholes rather t…
A paper from UC Berkeley, TTIC, and Google Research authored by Keegan Harris, Siddharth Prasad, Asher Trockman, Nika Haghtalab, and Michael I. Jordan shows that sorting an LLM inference queue strictl…
The Chinese Academy of Sciences' Institute of Artificial Intelligence for Industries released Maxwell, an embodied AI model that scored 91.9 on the Meta-World benchmark for robot physical tasks, the h…
A developer adapted Decision Transformers — the sequence-modeling reinforcement learning architecture from Chen et al. at Berkeley — to intelligent tutoring for heritage languages with fewer than 2,00…
Researchers are building AI-SQL query engines that push LLM inference into the database, but executing AI functions remains extremely expensive because a single query can trigger hundreds of thousands…
PrismML, an AI lab founded by Caltech researchers and advised by UC Berkeley's Ion Stoica, unveiled a 1-bit Bonsai LLM that runs locally on AI smart glasses built on Qualcomm's Snapdragon AR1 Gen 1 Pl…
A comparative review evaluates seven auto routing platforms for LLM model selection in 2026, including Bifrost, an open-source AI gateway written in Go that provides multi-provider routing, load balan…
Anthropic researchers found that a multi-agent research system with a lead agent dispatching parallel subagents beat a single agent by 90.2% on an internal eval, but consumed roughly 15 times the toke…
A test of a small prompt edit on 100 single-call tool-calling cases from UC Berkeley's BFCL V4 benchmark found the edit reduced invented optional arguments from 14 to 7 across three runs but produced …
MIT CSAIL director Daniela Rus and UC Berkeley professor Ion Stoica discussed infrastructure, memory, edge computing, and alignment challenges facing agentic AI, according to Forbes contributor John W…
A developer-written analysis attributes the widespread perception that large language models like Gemini and ChatGPT are getting worse to a mix of psychological and technical factors, including fading…
The Berkeley Overmind, an AI agent built by a UC Berkeley team led by computer science professor Dan Klein, defeated human StarCraft expert Oriol Vinyals for the first time after dozens of test matche…
Princeton's Pravesh Kothari, UCLA's Raghu Meka, and UC Berkeley's Prasad Raghavendra proposed a separate "conceptual track" for theoretical computer science conferences including STOC, FOCS, and SODA …
UC Berkeley researchers and the NVIDIA robotics team have developed Agentic Robotics (AR), an approach that uses multi-agent AI systems to write, test, and iteratively improve modular robot programs l…
Computing pioneer David Patterson, a former UC Berkeley computer science professor and now a distinguished engineer and fellow at Google, said at the AI Infra Summit that Jensen Huang is "unquestionab…
Researchers at UC Berkeley and Arena published HarnessTax on September 16, 2026, finding that swapping the harness around the same coding model changed task success by only about ±2 points on SWE-benc…
The San Mateo County Board of Supervisors voted unanimously Tuesday to create a first-of-its-kind permitting system for businesses using autonomous AI-controlled robots, effective mid-October and appl…
A failure taxonomy built from more than 1,600 annotated traces across seven multi-agent frameworks including AutoGen, CrewAI, LangGraph and OpenAI Swarm found that roughly 79% of multi-agent failures …
Salesforce is pitching its Trust Boundary as a unified architecture for agentic enterprises at Dreamforce 2026, but the agent trust stack remains split across three incompatible domains: governance sp…
A Prime Residency researcher ported the 105-task Linux CLI subset of Agents' Last Exam, a frontier benchmark from Berkeley RDI spanning 41 subdomains across 12 domain categories, to Prime Intellect's …