Unverified50% confidenceBenchmarkExact time
In the Artificial Analysis benchmark, GLM 5.3 Flash has a hallucination score of 7 and Qwen 3.8 (27B) scores -10, with local models having higher hallucination rates than cloud-based ones.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/10/2026
First Seen
Valid until: 12/9/2026
Sources
Related Entities
Related Claims
UnverifiedA calibration reprojection error of 1.8 pixels is 3 to 5 times higher than the industry quality standard (0.3–0.5 px) and will be amplified through PnP solving into centimeter-level 3D measurement deviations.62% similarUnverified光流估计（通过 RAFT、GMFlow 等算法）用于量化视频运动信息，光流均值接近零表示静态场景，方差过大意味着剧烈场景切换或抖动，这类数据不适合训练59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/887803API
curl https://kongchang.com/api/v1/knowledge/claims/887803MCP
get_claim(id=887803)