UK AISI uses EvalEval to share reproducible LLM benchmark results
uk aisi shares evaluation results for six frontier models using evaleval's infrastructure, promoting reproducible and verifiable evaluation science.
the uk ai security institute (aisi) has partnered with evaleval to share evaluation results for six frontier models: claude opus 4, claude opus 4.5, claude opus 4.6, gpt-5, gpt-5.2, and gpt-5.4. aisi is using evaleval's evaluation cards platform to provide transparent and reproducible evaluation results, including verified results, context, and configuration information for five benchmarks. this collaboration builds on previous research and aims to improve the reproducibility and verifiability of evaluation science. by sharing evaluation results and infrastructure, aisi and evaleval are working to diagnose gaps in evaluation reporting and build shared infrastructure to close them.