ReThinker: Scientific Reasoning by Rethinking with Guided Reflection and Confidence Control
About
Expert-level scientific reasoning remains challenging for large language models, particularly on benchmarks such as Humanity's Last Exam (HLE), where rigid tool pipelines, brittle multi-agent coordination, and inefficient test-time scaling often limit performance. We introduce ReThinker, a confidence-aware agentic framework that orchestrates retrieval, tool use, and multi-agent reasoning through a stage-wise Solver-Critic-Selector architecture. Rather than following a fixed pipeline, ReThinker dynamically allocates computation based on model confidence, enabling adaptive tool invocation, guided multi-dimensional reflection, and robust confidence-weighted selection. To support scalable training without human annotation, we further propose a reverse data synthesis pipeline and an adaptive trajectory recycling strategy that transform successful reasoning traces into high-quality supervision. Experiments on HLE, GAIA, and XBench demonstrate that ReThinker consistently outperforms state-of-the-art foundation models with tools and existing deep research systems, achieving state-of-the-art results on expert-level reasoning tasks.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Expert-Level Reasoning | HLE (Humanity's Last Exam) text-only subset (val) | Inference Accuracy52.2 | 13 | |
| Expert-Level Reasoning | GAIA text-only (val) | Inference Accuracy81.6 | 12 | |
| Expert-Level Reasoning | XBench-DeepSearch 1.0 (test) | Inference Accuracy0.9 | 12 |