---
format: "aidr-story-markdown/v1"
id: "8b1cc7e586f3f1bf2eecac9ddbc24b74c24002469766477a12fc1856f9061702"
canonical_url: "https://aidr.today/8b1cc7e5?lang=en"
title: "AI agents have bad taste in long-term decisions and a new benchmark proves it"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-10-18T08:49:41.000Z"
category: "Agents"
topics: ["openai","agent"]
source_urls: ["https://the-decoder.com/ai-agents-have-bad-taste-in-long-term-decisions-and-a-new-benchmark-proves-it/"]
summary: "A new benchmark called Taste-Bench evaluates how well AI agents make strategic choices during long-running tasks. The best model, GPT-5.6 Sol, selects the correct path on just under 60 percent of simple binary questions, and extended thinking does not improve performance."
---

# AI agents have bad taste in long\-term decisions and a new benchmark proves it

> [Open the canonical story](<https://aidr.today/8b1cc7e5?lang=en>)

**Published:** 2026-10-18T08:49:41.000Z
**Category:** Agents
**Topics:** openai, agent

## Summary

A new benchmark called Taste\-Bench evaluates how well AI agents make strategic choices during long\-running tasks\. The best model, GPT\-5\.6 Sol, selects the correct path on just under 60 percent of simple binary questions, and extended thinking does not improve performance\.

## Sources

- [Story source](<https://the-decoder.com/ai-agents-have-bad-taste-in-long-term-decisions-and-a-new-benchmark-proves-it/>)