Inference-Time Scaling for Complex Tasks: Where We Stand and What Lies Ahead
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915221170290688 |
|---|---|
| author | Balachandran, Vidhisha Chen, Jingya Chen, Lingjiao Garg, Shivam Joshi, Neel Lara, Yash Langford, John Nushi, Besmira Vineet, Vibhav Wu, Yue Yousefi, Safoora |
| author_facet | Balachandran, Vidhisha Chen, Jingya Chen, Lingjiao Garg, Shivam Joshi, Neel Lara, Yash Langford, John Nushi, Besmira Vineet, Vibhav Wu, Yue Yousefi, Safoora |
| contents | Inference-time scaling can enhance the reasoning capabilities of large language models (LLMs) on complex problems that benefit from step-by-step problem solving. Although lengthening generated scratchpads has proven effective for mathematical tasks, the broader impact of this approach on other tasks remains less clear. In this work, we investigate the benefits and limitations of scaling methods across nine state-of-the-art models and eight challenging tasks, including math and STEM reasoning, calendar planning, NP-hard problems, navigation, and spatial reasoning. We compare conventional models (e.g., GPT-4o) with models fine-tuned for inference-time scaling (e.g., o1) through evaluation protocols that involve repeated model calls, either independently or sequentially with feedback. These evaluations approximate lower and upper performance bounds and potential for future performance improvements for each model, whether through enhanced training or multi-model inference systems. Our extensive empirical analysis reveals that the advantages of inference-time scaling vary across tasks and diminish as problem complexity increases. In addition, simply using more tokens does not necessarily translate to higher accuracy in these challenging regimes. Results from multiple independent runs with conventional models using perfect verifiers show that, for some tasks, these models can achieve performance close to the average performance of today's most advanced reasoning models. However, for other tasks, a significant performance gap remains, even in very high scaling regimes. Encouragingly, all models demonstrate significant gains when inference is further scaled with perfect verifiers or strong feedback, suggesting ample potential for future improvements. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_00294 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Inference-Time Scaling for Complex Tasks: Where We Stand and What Lies Ahead Balachandran, Vidhisha Chen, Jingya Chen, Lingjiao Garg, Shivam Joshi, Neel Lara, Yash Langford, John Nushi, Besmira Vineet, Vibhav Wu, Yue Yousefi, Safoora Machine Learning Artificial Intelligence Computation and Language I.2 Inference-time scaling can enhance the reasoning capabilities of large language models (LLMs) on complex problems that benefit from step-by-step problem solving. Although lengthening generated scratchpads has proven effective for mathematical tasks, the broader impact of this approach on other tasks remains less clear. In this work, we investigate the benefits and limitations of scaling methods across nine state-of-the-art models and eight challenging tasks, including math and STEM reasoning, calendar planning, NP-hard problems, navigation, and spatial reasoning. We compare conventional models (e.g., GPT-4o) with models fine-tuned for inference-time scaling (e.g., o1) through evaluation protocols that involve repeated model calls, either independently or sequentially with feedback. These evaluations approximate lower and upper performance bounds and potential for future performance improvements for each model, whether through enhanced training or multi-model inference systems. Our extensive empirical analysis reveals that the advantages of inference-time scaling vary across tasks and diminish as problem complexity increases. In addition, simply using more tokens does not necessarily translate to higher accuracy in these challenging regimes. Results from multiple independent runs with conventional models using perfect verifiers show that, for some tasks, these models can achieve performance close to the average performance of today's most advanced reasoning models. However, for other tasks, a significant performance gap remains, even in very high scaling regimes. Encouragingly, all models demonstrate significant gains when inference is further scaled with perfect verifiers or strong feedback, suggesting ample potential for future improvements. |
| title | Inference-Time Scaling for Complex Tasks: Where We Stand and What Lies Ahead |
| topic | Machine Learning Artificial Intelligence Computation and Language I.2 |
| url | https://arxiv.org/abs/2504.00294 |