OSWorld
PulseAugur coverage of OSWorld — every cluster mentioning OSWorld across labs, papers, and developer communities, ranked by signal.
- used by Claude Haiku 5.5 70%
- instance of ScienceCast 70%
- used by Gotit.pub 70%
- instance of Gotit.pub 70%
- used by Influence Flower 70%
- instance of alphaXiv 70%
- used by CORE Recommender 70%
- used by ScienceCast 70%
- instance of DagsHub 60%
- used by alphaXiv 60%
- instance of CatalyzeX 60%
- used by DagsHub 60%
4 day(s) with sentiment data
-
Anthropic's Claude Sonnet 4.6 offers 1M context, Opus-level coding
Anthropic has released Claude Sonnet 4.6, a significant upgrade to its Sonnet model line. This new version boasts a 1 million token context window, enabling it to process extensive codebases or lengthy documents in a si…
-
New AI methods optimize visual token processing for efficiency
Researchers have developed new methods for optimizing visual token processing in AI models to reduce computational costs. Visual Token Routing (VTR) learns to select only the most relevant visual tokens from screenshots…
-
Anthropic's Claude Haiku 5.5 shows major gains in complex tasks
Anthropic has released Claude Haiku 5.5, a new iteration of its smaller language model, with significant performance improvements over its predecessor, Haiku 4.5. The model demonstrates a substantial leap in handling co…
-
Top LLMs for Browser Agents in 2026 Ranked by Score and Cost
A comparative analysis of LLMs for browser agents in October 2026 highlights Claude Opus 5.5 as the top performer for success rate, with Claude Sonnet 5.5 offering a similar score at half the price. For budget-conscious…
-
ReVision method slashes token costs for AI agents by 46%
Researchers have developed a new method called ReVision to improve the efficiency of computer-use agents (CUAs) by reducing the number of visual tokens they process. CUAs typically encode screenshots into many tokens, w…
-
New GUIFlow framework speeds up AI agents interacting with evolving interfaces
Researchers have developed GUIFlow, a novel framework designed to improve the efficiency of GUI agents that interact with evolving interfaces. By employing incremental inference, GUIFlow selectively updates computationa…
-
Anthropic's Claude Haiku 5.5 cuts prices but adds a 100K token cliff
Anthropic has released Claude Haiku 5.5, a new iteration of its cost-effective language model. While the base price for prompts under 100,000 tokens has significantly decreased, potentially by up to 90% compared to Haik…
-
AutoGUIWorld framework synthesizes GUI interaction data for AI agents
Researchers have developed AutoGUIWorld, a novel framework that synthesizes GUI interaction trajectories for training AI agents. This system leverages image generators and planners to create realistic interaction data w…
-
H Company releases Holo4 AI agent models for cross-platform use
H Company has launched Holo4, a new family of open-weight models designed for AI agents that can interact with desktops, web interfaces, and APIs. The models come in two sizes, Holo4 27B and Holo4 35B-A3B, offering a 25…
-
New CUA-Universe environment trains AI agents for hybrid GUI+CLI tasks
Researchers have developed CUA-Universe, a novel environment designed to train computer-use agents that can effectively combine graphical user interface (GUI) and command-line interface (CLI) interactions. This system a…
-
New CURA system detects AI agent failures with certified runtime alarms
A new system called CURA has been developed to monitor computer-use agents (CUAs) and detect failures in real-time. CURA operates by analyzing telemetry data, without needing access to the agent's internal workings or r…
-
New ASIL interface enhances AI agent interaction with software
Researchers have introduced ASIL (Agent-Software Interaction Layer), a new interface designed to improve how AI agents interact with software applications. ASIL replaces inefficient screenshot-and-click methods with str…
-
ChainWorld framework tests AI agents on complex, multi-step desktop tasks
A new research paper introduces ChainWorld, a framework designed to evaluate AI agents on complex, long-horizon desktop tasks. Unlike previous evaluations focused on single, atomic actions, ChainWorld composes these ato…
-
Qwen3.8-27B open-source model tops leaderboards with advanced agency
Alibaba's Qwen team has open-sourced Qwen3.8-27B, a 27-billion parameter model that has achieved top rankings on several benchmarks, including SWE Bench Pro and OSWorld. This model demonstrates significant advancements …
-
AI agents: Debugging frameworks offer repair over regeneration
Two research papers propose novel debugging frameworks for AI agents, shifting focus from regeneration to repair. The first, CUADebug, targets failures in computer-use agents by analyzing visual perception, spatial grou…
-
Chinese AI Agent InAgent Sets New Record on OSWorld Benchmark
InAgent, a Chinese AI agent, has achieved a record 90.2% success rate on the OSWorld benchmark, making it the first agent to surpass the 90% threshold. This performance exceeds previous records set by major AI labs like…
-
2026 LLM Benchmark: No Single Winner, Specialized Leaders Emerge · 1 source tracked
A comprehensive benchmark of 20 leading LLMs in 2026 reveals no single dominant model, but rather specialized leaders across different tasks. Claude Opus 5 leads the overall Artificial Analysis Intelligence Index, while…
-
New Interactive Reward Agent Achieves 86.9% Accuracy in GUI Task Evaluation
Researchers have developed an Interactive Reward Agent (IRA) designed to improve the evaluation of GUI agents. This agent utilizes a propose-then-verify framework to gather and confirm evidence from the environment's st…
-
AI agents struggle with visual tasks; EvoCUA-1.5 hits 63.2% on OSWorld benchmark
A new benchmark called OSWorld has revealed that the AI agent EvoCUA-1.5 achieved a score of 63.2% by learning through trial-and-error in simulated environments. Separately, a new open benchmark developed by Apple Inc. …
-
New BREW framework enables LLM agents to learn from experience
Researchers have developed BREW, a novel framework designed to enable Large Language Model (LLM)-based agents to learn from past experiences. Unlike current agents that restart learning with each session, BREW distills …