Expired50% confidenceFactExact time
LiveBench uses a continuously updated question bank to prevent models from gaming the benchmark.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/19/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/17/2026(expired)
Sources
GPT-5.6发布即预警：算力瓶颈才是大模型真正的考验
Related Claims
UnverifiedLiveBench uses a continuously updated question pool, replacing some questions monthly to combat data contamination.80% similarUnverifiedModel providers frequently perform silent updates to adjust model weights, safety filters, or system-level prompts, but often without explicit changelogs, leaving users unable to track changes or roll back.58% similarUnverifiedprops-for-that provides live-type properties (continuously updated in real time, such as --live-visible toggling between 1 and 0) and const-type properties (triggered only once, such as --const-has-entered, which becomes 1 upon first entering the viewport and does not change again).55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/565006API
curl https://kongchang.com/api/v1/knowledge/claims/565006MCP
get_claim(id=565006)