Search papers, labs, and topics across Lattice.
This paper introduces Test-Time Tool Evolution (TTE), a paradigm where LLM agents synthesize, verify, and evolve executable tools during inference to address the limitations of static tool libraries in scientific reasoning. TTE overcomes the rigidity of predefined tools by treating them as problem-driven artifacts, enabling adaptation to sparse and heterogeneous scientific domains. Experiments on the newly introduced SciEvo benchmark, comprising 1,590 tasks and 925 evolved tools, demonstrate that TTE achieves state-of-the-art performance in accuracy, tool efficiency, and cross-domain adaptation.
LLMs can now dynamically create and refine their own scientific tools at test time, outperforming agents stuck with static toolsets.
The central challenge of AI for Science is not reasoning alone, but the ability to create computational methods in an open-ended scientific world. Existing LLM-based agents rely on static, pre-defined tool libraries, a paradigm that fundamentally fails in scientific domains where tools are sparse, heterogeneous, and intrinsically incomplete. In this paper, we propose Test-Time Tool Evolution (TTE), a new paradigm that enables agents to synthesize, verify, and evolve executable tools during inference. By transforming tools from fixed resources into problem-driven artifacts, TTE overcomes the rigidity and long-tail limitations of static tool libraries. To facilitate rigorous evaluation, we introduce SciEvo, a benchmark comprising 1,590 scientific reasoning tasks supported by 925 automatically evolved tools. Extensive experiments show that TTE achieves state-of-the-art performance in both accuracy and tool efficiency, while enabling effective cross-domain adaptation of computational tools. The code and benchmark have been released at https://github.com/lujiaxuan0520/Test-Time-Tool-Evol.
