Research
When More Thinking Hurts: Overthinking in LLM Test-Time Compute Scaling
arXiv:2604.10739v1 Announce Type: new Abstract: Scaling test-time compute through extended chains of thought has become a dominant paradigm for improving large language model reasoning. However, exist
arXiv:2604.10739v1 Announce Type: new Abstract: Scaling test-time compute through extended chains of thought has become a dominant paradigm for improving large language model reasoning. However, existing research implicitly assumes that longer thinking always yields better results. This assumption remains largely unexamined. We systematically investigate how the marginal utility of additional reasoning tokens changes as compute budgets increase. We find that marginal returns diminish substantially at higher budgets and that models exhibit ``overthinking'', where extended reasoning is associated with abandoning previously correct answers. Furthermore, we show that optimal thinking length varies across problem difficulty, suggesting that uniform compute allocation is suboptimal. Our cost-aware evaluation framework reveals that stopping at moderate budgets can reduce computation significantly while maintaining comparable accuracy.
Related
- Reasoning Fails Where Step Flow Breaks
- STU-PID: Steering Token Usage via PID Controller for Efficient Large Language Model Reasoning
- On the Step Length Confounding in LLM Reasoning Data Selection
- Rectifying LLM Thought from Lens of Optimization
- TokUR: Token-Level Uncertainty Estimation for Large Language Model Reasoning
Source: arXiv cs.AI | 2026-04-14