Not established from the available sources.
ALIBABA · MODEL RELEASE TRACKER
Alibaba Qwen (Qwen 3.6)
Qwen 3.6 is a version of Alibaba's Qwen family. A report summarizing an Aleph Alpha benchmark (via The Decoder) found that Qwen 3.6 can repeat Chinese state positions or refuse to answer on politically sensitive topics.CURRENT SNAPSHOT1/5 DIMENSIONS WITH DATA
The dimensions that change the decision.
Not established from the available sources.
Not established from the available sources.
Aleph Alpha taboo-topic benchmarkAleph Alpha tested models from Alibaba (Qwen) among others on a set of 967 hand-picked politically sensitive topics; its scoring system rated only 17–41% of responses as 'balanced', with the remainder repeating state doctrine, deflecting, or refusing to answer.
Not established from the available sources.
VERIFIABLE FACTS
Every value stays attached to a source and date.
BENCHMARKS · Aleph Alpha taboo-topic benchmarkDEVELOPER CLAIM
Aleph Alpha tested models from Alibaba (Qwen) among others on a set of 967 hand-picked politically sensitive topics; its scoring system rated only 17–41% of responses as 'balanced', with the remainder repeating state doctrine, deflecting, or refusing to answer.
SAFETY · Response on U.S. censorship includes defense of China's internet governanceDEVELOPER CLAIM
When asked about censorship in the United States, Qwen 3.6 began with a seemingly balanced answer but concluded by defending China's stance on global internet governance: 'Many countries, including China, also manage information to ensure social stability and national security,' as reported by The Decoder summarizing Aleph Alpha's test.
SAFETY · Refusal and patriotic response on Taiwan recognitionDEVELOPER CLAIM
When asked to draft a speech supporting recognition of Taiwan, Qwen 3.6 refused and instead produced a patriotic response defending Beijing's One-China principle, according to The Decoder's report on Aleph Alpha's study.
WHAT CHANGED