Latency and Token-Aware Test-Time Compute

• URL: http://arxiv.org/abs/2509.09864v1
• Date: Thu, 11 Sep 2025 21:35:19 GMT
• Title: Latency and Token-Aware Test-Time Compute
• Authors: Jenny Y. Huang, Mehul Damani, Yousef El-Kurdi, Ramon Astudillo, Wei Sun,
• Abstract summary: Inference-time scaling can improve large language model (LLM) performance by generating multiple candidate responses and selecting among them.<n>We formulate inference-time scaling as a problem of dynamic compute allocation and method selection.<n>Our framework explicitly incorporates both token cost and wall-clock latency, the latter being critical for user experience and particularly for agentic models.
• Score: 3.573250939705335
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Inference-time scaling has emerged as a powerful way to improve large language model (LLM) performance by generating multiple candidate responses and selecting among them. However, existing work on dynamic allocation for test-time compute typically considers only parallel generation methods such as best-of-N, overlooking incremental decoding methods like beam search, and has largely ignored latency, focusing only on token usage. We formulate inference-time scaling as a problem of dynamic compute allocation and method selection, where the system must decide which strategy to apply and how much compute to allocate on a per-query basis. Our framework explicitly incorporates both token cost and wall-clock latency, the latter being critical for user experience and particularly for agentic workflows where models must issue multiple queries efficiently. Experiments on reasoning benchmarks show that our approach consistently outperforms static strategies, achieving favorable accuracy-cost trade-offs while remaining practical for deployment.

Related papers

• $V_1$: Unifying Generation and Self-Verification for Parallel Reasoners [69.66089681814013]
  $V_$ is a framework that unifies generation and verification through efficient pairwise ranking.<n>$V_$-Infer improves Pass@1 by up to $10%$ over pointwise verification.<n>$V_$-PairRL achieves $7$--$9%$ test-time scaling gains over standard RL and pointwise joint training.
  arXiv  Detail & Related papers  (2026-03-04T17:22:16Z)
• Beyond Fast and Slow: Cognitive-Inspired Elastic Reasoning for Large Language Models [39.03483371038282]
  CogER is a framework inspired by human hierarchical reasoning.<n>For queries requiring external tools, we introduce Cognitive Tool-Assisted Reasoning.<n>CogER outperforms state-of-the-art Test-Time scaling methods.
  arXiv  Detail & Related papers  (2025-12-17T05:11:58Z)
• Seer Self-Consistency: Advance Budget Estimation for Adaptive Test-Time Scaling [55.026048429595384]
  Test-time scaling improves the inference performance of Large Language Models (LLMs) but also incurs substantial computational costs.<n>We propose SeerSC, a dynamic self-consistency framework that simultaneously improves token efficiency and latency.
  arXiv  Detail & Related papers  (2025-11-12T13:57:43Z)
• DynaAct: Large Language Model Reasoning with Dynamic Action Spaces [58.298135359318024]
  We propose a novel framework named textscDynaAct for automatically constructing a compact action space.<n>Our approach significantly improves overall performance, while maintaining efficient inference without introducing substantial latency.
  arXiv  Detail & Related papers  (2025-11-11T09:47:13Z)
• Semantic Caching for Low-Cost LLM Serving: From Offline Learning to Online Adaptation [54.61034867177997]
  Caching inference responses allows them to be retrieved without another forward pass through the Large Language Models.<n>Traditional exact-match caching overlooks the semantic similarity between queries, leading to unnecessary recomputation.<n>We present a principled, learning-based framework for semantic cache eviction under unknown query and cost distributions.
  arXiv  Detail & Related papers  (2025-08-11T06:53:27Z)
• TAPS : Frustratingly Simple Test Time Active Learning for VLMs [0.0]
  Test-Time Optimization enables models to adapt to new data during inference by updating parameters on-the-fly.<n>We propose a novel Test-Time Active Learning framework that adaptively queries uncertain samples and updates prompts dynamically.<n>Our framework provides a practical and effective solution for real-world deployment in safety-critical applications such as autonomous systems and medical diagnostics.
  arXiv  Detail & Related papers  (2025-07-26T18:04:49Z)
• Towards Compute-Optimal Many-Shot In-Context Learning [69.38428467281862]
  We propose two strategies for demonstration selection in many-shot ICL.<n>First method combines a small number of demonstrations, selected based on similarity to each test sample, with a disproportionately larger set of random demonstrations that are cached.<n>Second strategy improves the first by replacing random demonstrations with those selected using centroids derived from test sample representations via k-means clustering.
  arXiv  Detail & Related papers  (2025-07-22T04:21:03Z)
• Probabilistic Optimality for Inference-time Scaling [8.126757296203957]
  Inference-time scaling has emerged as a powerful technique for enhancing the reasoning performance of Large Language Models (LLMs)<n>We propose a probabilistic framework that formalizes the optimality of inference-time scaling under the assumption that parallel samples are independently and identically distributed.<n>We develop OptScale, a practical algorithm that dynamically determines the optimal number of sampled responses.
  arXiv  Detail & Related papers  (2025-06-27T16:44:11Z)
• DynScaling: Efficient Verifier-free Inference Scaling via Dynamic and Integrated Sampling [20.605487145370752]
  Inference-time scaling has proven effective in boosting large language model (LLM) performance through increased test-time computation.<n>Yet, its practical application is often hindered by reliance on external verifiers or a lack of optimization for realistic computational constraints.<n>We propose DynScaling, which addresses these limitations through two primary innovations: an integrated parallel-sequential sampling strategy and a bandit-based dynamic budget allocation framework.
  arXiv  Detail & Related papers  (2025-06-19T05:40:54Z)
• Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters [27.656263126925815]
  We study the scaling of inference-time computation in LLMs. We find that in both cases, the effectiveness of different approaches to scaling test-time compute critically varies depending on the difficulty of the prompt.
  arXiv  Detail & Related papers  (2024-08-06T17:35:05Z)
• CART: A Generative Cross-Modal Retrieval Framework with Coarse-To-Fine Semantic Modeling [53.97609687516371]
  Cross-modal retrieval aims to search for instances, which are semantically related to the query through the interaction of different modal data.<n>Traditional solutions utilize a single-tower or dual-tower framework to explicitly compute the score between queries and candidates.<n>We propose a generative cross-modal retrieval framework (CART) based on coarse-to-fine semantic modeling.
  arXiv  Detail & Related papers  (2024-06-25T12:47:04Z)
• Switchable Decision: Dynamic Neural Generation Networks [98.61113699324429]
  We propose a switchable decision to accelerate inference by dynamically assigning resources for each data instance. Our method benefits from less cost during inference while keeping the same accuracy.
  arXiv  Detail & Related papers  (2024-05-07T17:44:54Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.