Expired50% confidenceSolutionExact time
For everyday tasks, low or medium level (IQ approximately 105–120) is sufficient; extra high (IQ up to 135) should only be considered when handling complex logic.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/18/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/16/2026(expired)
Sources
Related Claims
Unverified领域微调后的7B模型在任务准确率上往往超越未经微调的70B通用模型，同时推理成本降低一个数量级66% similarUnverifiedThe Comprehensive Intelligence Index consists of 9 independent evaluations including agentic work, coding, scientific reasoning, and general capability; Sol and Faber-5 are nearly on par with a difference of only 1 point, but task completion time was reduced by approximately 61% and cost was approximately half.65% similarUnverified根据文中量化估算，编程智能体任务中模型推理仅占总时间的 20%-30%，剩余 70%-80% 花在等待外部系统 I/O 响应64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/552245API
curl https://kongchang.com/api/v1/knowledge/claims/552245MCP
get_claim(id=552245)