Unverified50% confidenceBenchmarkExact time
The study found that GPT-4's accuracy on the prime number identification task plummeted from 97.6% in March 2023 to 2.4% in June 2023.
1
Sources
50%
Confidence
Long-term
Relevance
8/31/2026
First Seen
Sources
Related Entities
Related Claims
Unverified在GPT-4发布初期，主流模型在GPQA上的准确率约为35-40%，人类领域专家基准线约为65%，随机猜测为25%79% similarUnverifiedGPTQ（2022年）和AWQ（2023年）能将70B参数的LLM量化到INT4精度，困惑度损失不到1%70% similarUnverifiedEarly GPT-4 had a pass rate of only 1.7% on SWE-bench, Claude 3 Opus approximately 4.8%, and it was not until mid-2024 that the 20% mark was surpassed.70% similarUnverifiedIn SWE-bench testing, the early GPT-4 resolution rate was below 2%, while Claude 3.5 Sonnet broke through 49%.69% similarUnverifiedIn the Agents' Last Exam benchmark, GPT-5.6 Sol achieved a score of 53.6, which is 13.1 points higher than Claude Fable 5 using adaptive reasoning.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/830050API
curl https://kongchang.com/api/v1/knowledge/claims/830050MCP
get_claim(id=830050)