Expired50% confidenceBenchmarkExact time
According to public evaluations, the success rate of existing desktop agents in complex workflows is generally below 50%.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/20/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/18/2026(expired)
Sources
Related Claims
UnverifiedEven GPT-4-level Agents often achieve end-to-end success rates below 60% on complex tasks involving more than 5 tool calls.69% similarUnverifiedFor a task that sequentially calls 5 APIs, even if the success rate of each step is 95%, the overall success rate drops to approximately 77%.68% similarUnverifiedWhen the number of chain nodes exceeds 5, the overall success rate of Agent systems is often less than 60%.68% similarVerifiedCurrently, the resolution rate of top models on the SWE-bench Verified subset has exceeded 50%.68% similarUnverified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/575454API
curl https://kongchang.com/api/v1/knowledge/claims/575454MCP
get_claim(id=575454)