Unverified50% confidenceBenchmarkExact time
When cues appeared concentrated, GPT-5.5, Grok 4.3, DeepSeek V4 and Llama 4 Maverick met the support workers' gold standard in only 26% to 59% of conversations.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
11/20/2026
First Seen
Valid until: 12/4/2026
Sources
Related Entities
Related Claims
UnverifiedGPT-5.4 was reversed by a "trusted source" on 44.7% of questions, Grok-4.20 had a reversal rate as high as 87.5%, while Gemini-3.1-Pro was nearly unaffected at just 0.6%.71% similarVerifiedMultiple academic studies have shown that a considerable proportion of the high scores achieved by mainstream models such as GPT-4 and LLaMA on leaderboards including MMLU, HumanEval, and GSM8K stem from overlap between training set and test set content.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1190480API
curl https://kongchang.com/api/v1/knowledge/claims/1190480MCP
get_claim(id=1190480)