Unverified50% confidenceBenchmarkExact time
In multiple benchmarks, even the most advanced models can have hallucination rates of 15%–30% in specialized technical domains.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/2/2026
First Seen
Valid until: 12/1/2026
Sources
Related Entities
Related Claims
Unverified早期模型在某些领域的事实幻觉率高达15-25%，尤其在长尾知识领域幻觉率更高75% similarUnverifiedStanford research shows that in complex debugging scenarios, the probability of LLM suggestions causing beginners to form incorrect mental models is approximately 23%, far higher than that of human tutors.73% similarVerifiedResearch shows that even the most advanced models have hallucination rates in factual question-answering tasks that still fluctuate between 5% and 20%.71% similarUnverifiedResearch shows that hallucination issues are particularly prominent in specialized fields such as healthcare, law, and finance, where error rates can be as high as 30% or more.71% similarUnverifiedRöttger et al. found in a systematic evaluation in 2024 that the over-refusal rate of mainstream models exceeded 30% in certain categories.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/843240API
curl https://kongchang.com/api/v1/knowledge/claims/843240MCP
get_claim(id=843240)