Unverified50% confidenceBenchmarkExact time
GSM8K发布时GPT-3的zero-shot准确率仅约5%，加入few-shot示例也不超过35%
1
Sources
50%
Confidence
Long-term
Relevance
7/17/2026
First Seen
Sources
Related Claims
Unverified在 GSM8K 基准上，标准思维链准确率约为 56%-63%（GPT-3 规模），PAL 可提升至 72% 以上70% similarUnverifiedIn SWE-bench testing, the early GPT-4 resolution rate was below 2%, while Claude 3.5 Sonnet broke through 49%.67% similarUnverifiedIn the initial publicly released SWE-bench results, GPT-4's resolution rate was only approximately 1.7%, Devin claimed to reach 13.86%, and SWE-agent reached 12.5%.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/542332API
curl https://kongchang.com/api/v1/knowledge/claims/542332MCP
get_claim(id=542332)