SemiAnalysis: only 3.6% of Chinese AI models have disclosed safety evaluation results, 1.1% before launch (report)

What happened

According to an AI Times report, AI analysis firm SemiAnalysis published an analysis on the 9th (local time) of how Chinese AI developers disclose model safety evaluations. The analysis covered models released from 2021 through September 15, 2026 by Alibaba, ByteDance, Tencent, Baidu, DeepSeek, Moonshot AI, Z.ai and others. Only 1.1% of models had safety verification results disclosed at launch, and even including cases where some results were disclosed well after release, the share was just 3.6%.

Provider claims

SemiAnalysis interpreted the findings as showing that although the Chinese government officially acknowledges AI safety risks, the industry in practice prioritizes model performance and capturing market share (as quoted in the report).

Why it matters

More companies are adopting Chinese open-weight models, but public evidence of these models' safety verification is almost nonexistent. This increases the burden on adopters to run safety checks such as red-team evaluations themselves.

Sources

• Press 중국 AI 모델 안전성 평가 공개율 3.6%...출시 전 공개는 고작 1.1% AI Times (Korea)

More about SemiAnalysis

• 2026-10-08 P2 Other Unconfirmed Language modelsData & evaluation SemiAnalysisAnthropicOpenAI
  SemiAnalysis: on the same $200/month plan, Claude Opus 5.5 usage value is over 5x that of GPT-6.1 Sol (report)

  According to AI Times, SemiAnalysis on the 6th (local time) published its own measurements of the actual usage limits of major AI...