---
format: "aidr-story-markdown/v1"
id: "08cebc8f97d4c11cb0c2601762467e3a67f16f7ef81017a5818f1ffeb9d0ed52"
canonical_url: "https://aidr.today/08cebc8f?lang=en"
title: "Selection-Based Structured Reasoning: Toward Efficient Multimodal Search Agents"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-10-15T00:00:00.000Z"
category: "Research"
topics: ["reasoning"]
source_urls: ["https://machinelearning.apple.com/research/selection-structured-reasoning"]
summary: "Selection-Based Structured Reasoning (SSR) is a framework that reformulates reasoning as selection from pre-specified natural-language candidates, achieving over 90% reduction in per-turn reasoning latency and 28–54% reduction in total inference latency on multimodal search benchmarks."
---

# Selection\-Based Structured Reasoning: Toward Efficient Multimodal Search Agents

> [Open the canonical story](<https://aidr.today/08cebc8f?lang=en>)

**Published:** 2026-10-15T00:00:00.000Z
**Category:** Research
**Topics:** reasoning

## Summary

Selection\-Based Structured Reasoning \(SSR\) is a framework that reformulates reasoning as selection from pre\-specified natural\-language candidates, achieving over 90% reduction in per\-turn reasoning latency and 28–54% reduction in total inference latency on multimodal search benchmarks\.

## Sources

- [Story source](<https://machinelearning.apple.com/research/selection-structured-reasoning>)