Unverified50% confidenceBenchmarkExact time

据公开评测现有桌面Agent在复杂工作流中的成功率普遍低于50%

1

Sources

50%

Confidence

Medium-term (~90 days)

Relevance

7/20/2026

First Seen

Valid until: 10/18/2026

Sources

OpenAI's Five-in-One Launch: Sol, Desktop Agent, and Real-Time Translation Explained

Related Claims

Unverified即便是GPT-4级别的Agent，面对含有5步以上工具调用的复杂任务时，端到端成功率往往低于60%69% similar Unverified当链路节点超过5个时，Agent系统整体成功率往往不足60%68% similar VerifiedCurrently, the resolution rate of top models on the SWE-bench Verified subset has exceeded 50%.68% similar Unverified推测解码中当草稿通过率低于50%时，实际加速效果会大打折扣甚至产生负优化68% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/575454

API

curl https://kongchang.com/api/v1/knowledge/claims/575454

MCP

get_claim(id=575454)