Unverified50% confidenceFactExact time
Text Arena uses anonymous head-to-head battles and blind human evaluations, ranking models through an Elo scoring mechanism.
1
Sources
50%
Confidence
Long-term
Relevance
8/6/2026
First Seen
Sources
阿里通义冲上Text Arena第二名，Qwen模型实力解析
Related Claims
UnverifiedArena-type platforms use a pairwise blind-testing mechanism in which users vote for the better response without knowing the identity of the models.65% similarUnverifiedWebDev Arena draws on the human preference voting mechanism of Chatbot Arena, ranking participants by ELO score through blind-test voting.65% similarUnverifiedChatbot Arena的ELO机制存在局限性，评判者主观偏好（如偏好更长回答）可能引入系统性偏差60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/699030API
curl https://kongchang.com/api/v1/knowledge/claims/699030MCP
get_claim(id=699030)