ResearchBenchmarksShort-query Safety on Short-query Safety English SuiteFollow100XSTest Safety ScoreOyster-I99.4899.61599.7599.885Jul 3, 2026Evaluation ResultsMethodMethodLinksXSTest Safety ScoreStrongReject RateDo Not Answer RateWildChat Safety ScoreOyster-I2026.0710010099.7992.69Oyster-II2026.0710010099.7997.25Qwen3-14B2026.0799.596.1799.6883.56