PhySciBench benchmark shows current AI models achieve at most 33.5% accuracy on physical science tasks; DelveAgent framework improves accuracy by up to 7.5 points and cuts costs to one-third.
CoRR , volume =
10 Pith papers cite this work. Polarity classification is still indexing.
10
Pith papers citing it
abstract
The rapid advancements of AI agents have ignited the long-held ambition of leveraging them to accelerate scientific discovery. Achieving this goal requires a deep understanding of the frontiers of human knowledge. As such, Humanity's Last Exam (HLE) provides an exceptionally challenging touchstone for evaluating scientific AI agents. In this work, we aim to construct the foundational architecture for general-purpose agents and validate the capabilities through leading performance on HLE. To achieve this, we introduce X-Master, a tool-augmented reasoning agent designed to emulate human researchers by interacting flexibly with external tools during its reasoning process. This agent, guided by the conceptualization of code as an interaction language, can flexibly leverage built-in Python libraries and our customized tools to augment the reasoning. We further scale its capabilities through X-Masters, a scattered-and-stacked agentic workflow that systematically enhances breadth and depth of reasoning. Our open-source solution, X-Masters, sets a new state-of-the-art record on HLE with a score of 32.1%, surpassing OpenAI's and Google's Deep Research (26.6% and 26.9%) and becoming the first to exceed the 30% threshold. This work allows us to gain a deeper understanding of complex task-solving and accumulates valuable experience that can inform future advancements, guiding subsequent model training.
citation-role summary
background 3
baseline 1
citation-polarity summary
years
2026 10representative citing papers
ReCrit frames critic interaction as a correctness-transition problem and uses quadrant-based RL rewards to improve LLM performance on scientific reasoning benchmarks by rewarding corrections and robustness while penalizing sycophancy.
DataPRM is an environment-aware generative process reward model that improves LLM data analysis agents by 7-11% on benchmarks via active verification and reflection-aware ternary rewards.
IntrAgent uses a two-stage pipeline of section ranking and iterative reading to perform content-grounded literature information retrieval, achieving 13.2% higher accuracy than RAG and agent baselines on the new IntraBench benchmark.
Arbor combines a coordinator, executors, and a hypothesis tree to enable cumulative autonomous research, outperforming Codex and Claude Code by over 2.5x on six real tasks and reaching 86.36% Any Medal on MLE-Bench Lite.
SciResearcher is a new agentic data-construction framework that trains an 8B model via supervised fine-tuning and reinforcement learning to reach 19.46% on HLE-Bio/Chem-Gold and 13-15% gains on related biology and literature benchmarks.
EvoMaster is an open-source agent framework whose iterative self-evolution loop achieves the highest average score on 9 of 10 scientific agent benchmarks compared to OpenHands, OpenClaw, and Codex.
SciAtlas builds a large-scale multi-disciplinary academic knowledge graph and a neuro-symbolic retrieval system to support automated scientific research tasks such as literature review and idea positioning.
The paper delivers a stage-by-stage roadmap for AI in research, showing reliable assistance in retrieval and tool tasks but fragility in novelty and judgment, advocating human-governed collaboration.
AI will evolve from a research tool into a collaborator, fundamentally reshaping scientific collaboration, discovery, publishing, and evaluation while requiring continuous learning and idea diversity for original contributions.
citing papers explorer
-
EvoMaster: A Foundational Evolving Agent Framework for Agentic Science at Scale
EvoMaster is an open-source agent framework whose iterative self-evolution loop achieves the highest average score on 9 of 10 scientific agent benchmarks compared to OpenHands, OpenClaw, and Codex.