Primary sourceGeek
View sourceEvent ·Multi-source
DeepSeek V4 Flash登顶NYT Connections新基准
在包含940道题的Extended NYT Connections基准上,DeepSeek V4 Flash取得89.6分,位列第一。Gemini 3.6 Flash以89.0分紧随其后。Qwen 3.7 Plus的成绩为74.8分,Gemini 3.5 Flash-Lite为60.4分。Qwen 3.7 Flash和Qwen3.6-35B-A3B分别只拿到43.8和41.6分。
Current read
DeepSeek V4 Flash在谜题基准拿到89.6,赢了Gemini 3.6的89.0,推理表现很亮眼。
11 sources67° AI heatLast updated 2026/8/2 05:05:40
Evidence
11 sourcesRelated source 1Google Gemini App
View sourceRelated source 2lmarena.ai
View sourceRelated source 3arXiv: DeepSeek
View sourceRelated source 4IT之家
View sourceRelated source 5深度求索 DeepSeek
View sourceRelated source 6歸藏(guizang.ai)
View sourceRelated source 7向阳乔木
View sourceRelated source 8orange.ai
View sourceRelated source 9AI Will
View sourceRelated source 10elvis
View sourceConflict check
Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.