Unverified50% confidenceFactExact time
Controversies over model capability degradation had previously emerged on a large scale among GPT-4 users, with some users claiming the model had become lazy or perfunctory.
1
Sources
50%
Confidence
Long-term
Relevance
8/31/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedGPT-5.6依然存在过度设计的倾向，需要人为纠偏避免把简单需求复杂化71% similarUnverified评分模型（如GPT-4）在版本更新后，其对优质回答的判断标准可能发生偏移，导致历史评分数据出现系统性断层71% similarUnverifiedLarge models such as GPT-4 can explain joke structures to a certain extent, but still frequently fail when facing real social contexts that require embodied experience71% similarUnverifiedThe industry widely believes that GPT-4 incorporates a Mixture of Experts (MoE) mechanism that dynamically routes and activates partial expert sub-networks during each inference69% similarUnverifiedA large number of users have reported that GPT-5.5 suffers from degraded output quality.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/828201API
curl https://kongchang.com/api/v1/knowledge/claims/828201MCP
get_claim(id=828201)