Attention @arxiv authors: Our Code of Conduct states that by signing your name as an author of a paper, each author takes full responsibility for all its contents, irrespective of how the contents were generated. 1/

Spotlighting our newest benchmark for agentic search: DETOUR When people try to recall something in conversation, they rarely give a perfect query upfront. They say things like “that movie with the scene where…” or “the paper about…” and the assistant has to ask the right follow-up questions to get there. Existing search and agent benchmarks often miss this multi-turn, tip-of-the-tongue behavior. To more realistically evaluate it, we introduce DETOUR: Dual-agent based Evaluation Through Obscure Under-specified Retrieval, an interactive benchmark for dual-agent search and reasoning. DETOUR contains 1,011 prompts across text, image, audio, and video. In the benchmark, a Primary Agent is evaluated on its ability to identify a target entity by querying a consistent Memory Agent, testing whether models can resolve ambiguity through useful follow-up questions. Current state-of-the-art models still struggle: performance reaches only 36% accuracy across all modalities, showing that today’s agents remain weak at clarification-seeking in underspecified, real-world search settings. We hope DETOUR helps push the next generation of search agents toward better reasoning, better questions, and more robust multi-turn retrieval. arXiv Paper: arxiv.org/abs/2602.00352

RL coding agents increasingly game rewards by exploiting their semantic and syntactic weaknesses. Can LLMs detect such behaviors from live training rollouts? We find contrastive cluster analysis is key! 🚀 GPT-5.2 jumps from 45% to 63%. Humans reach 90% Paper data 🧵

🚨We will be presenting Memtrack today at the SEA workshop from 3:50pm onwards at #NeurIPS2025 Memtrack is a SoTA eval env to study an agent's ability to memorize and retrieve facts using exploration over interleaved enterprise slack, linear and git threads in a multi-QA setting

if you or a loved one is looking to learn about building environments and get a bag in the process, inquire within our bounty list is bigger and better than ever

We’re excited to support @Meta and @huggingface's OpenEnv launch today! OpenEnv provides an open-source framework for building and interacting with agentic execution environments. This allows researchers and developers to create isolated, secure, deployable, and usable environments. Lately, at Patronus, we’ve been working on RL environments for coding agents, and we were excited to contribute to OpenEnv with real-world-inspired tools and tasks to train and steer AGI. We began with a Gitea-based git server environment. Git server environments are foundational and enable effective collaboration and version control for software workflows, and we thought it would be a perfect way to get started with OpenEnv. With our git server environment, we support: * Fast iteration across runs with sub-second resets for RL training loops * Shared server isolated workspaces * Environment variables setting custom configs for Gitea We look forward to seeing what everyone builds with OpenEnv! GitHub: github.com/meta-pytorch/Open… HuggingFace: huggingface.co/openenv

1/ 🔥🔥 Big news: We’re launching Percival, the first AI agent that can evaluate and fix other AI agents! 🤖 Percival is an evaluation agent that doesn’t just detect failures in agent traces — it can fix them. Percival outperformed SOTA LLMs by 2.9x on the TRAIL dataset, containing human annotated errors from GAIA and SWE-Bench. 🦾 Here’s what Percival can do for you: - Automatically suggest prompt fixes for your agent - Catch 20 types of agent failures spanning tool use, planning and coordination, domain specific errors - Reduce manual debugging time from hours to < 1 minute

Happy to see @PatronusAI use our Anchored Preference Optimization (APO) objective in their study!