Unverified50% confidenceFactExact time
The research team evaluated 14 combinations of models and reasoning modes.
1
Sources
50%
Confidence
Long-term
Relevance
10/22/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe researchers re-scored or re-generated results for 15 systems under a unified implementation framework.70% similarUnverifiedThe research team evaluated 18 proprietary and open-weight models on MindEval2.70% similarUnverifiedThe research evaluated multiple model configurations across different model families, different architectures, and different reasoning strategies.70% similarUnverifiedThe experiment ran 15 sessions covering two different model families, all using free-tier models.69% similarUnverifiedThe research team selected the Generalized Additive Model (GAM) as the classifier from six candidate models.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1084336API
curl https://kongchang.com/api/v1/knowledge/claims/1084336MCP
get_claim(id=1084336)