---
title: "Inference-Time Scaling (Test-Time Compute)"
description: "Inference-time scaling, also called test-time compute scaling, is the practice of allocating additional computation to a model after training in order to improve the quality of its answer. Instead of relying on a single forward generation, the system may let a reasoning model reason for longer, generate multiple candidates, search over possible solutions, or verify intermediate results."
canonical_url: "https://promptmetheus.com/resources/llm-knowledge-base/inference-time-scaling-test-time-compute"
---

> Index of every page on this site: [llms.txt](https://promptmetheus.com/llms.txt)

Inference-time scaling, also called test-time compute scaling, is the practice of allocating additional computation to a model after training in order to improve the quality of its answer. Instead of relying on a single forward generation, the system may let a [reasoning model](https://promptmetheus.com/resources/llm-knowledge-base/reasoning-model) reason for longer, generate multiple candidates, search over possible solutions, or verify intermediate results.

This differs from training-time scaling, where capability is increased by using more parameters, data, or training compute. Inference-time scaling changes the amount of work performed for each request and can therefore be adjusted dynamically according to task difficulty, latency requirements, and cost.

Common approaches include:

- producing several candidate answers and selecting one with a verifier;
- iterative self-correction or critique;
- tree or graph search over reasoning steps;
- executing code or tests to validate candidates;
- increasing a model's reasoning-effort budget; and
- using [tool calls](https://promptmetheus.com/resources/llm-knowledge-base/tool-calling-function-calling) to gather evidence before answering.

More compute does not automatically produce a better result. Repeated samples can share the same misconception, weak verifiers can select confidently incorrect answers, and long reasoning traces increase latency and [token](https://promptmetheus.com/resources/llm-knowledge-base/token) cost. Effective systems allocate compute selectively rather than applying the same budget to every request.

Inference-time scaling is especially relevant to [AI agents](https://promptmetheus.com/resources/llm-knowledge-base/ai-agent), which can spend additional steps planning, testing actions, and recovering from errors. It should be evaluated against end-to-end accuracy and resource usage through [agent evaluation](https://promptmetheus.com/resources/llm-knowledge-base/agent-evaluation).

The paper [Scaling LLM Test-Time Compute Optimally](https://arxiv.org/abs/2408.03314) shows that the most effective strategy depends on problem difficulty and that adaptively allocated test-time compute can outperform substantially larger models on suitable tasks.

## Sitemap

See the full [sitemap](https://promptmetheus.com/sitemap.md) for all pages.