CAIS AI Dashboard
dashboard.safe.ai · 158 words · saved by 6 readers
Evaluating frontier AI systems on capability and safety benchmarks
Evaluating frontier AI systems on capability and safety benchmarks Text Capabilities Index Size: StandardMini Model| GPT-6 Astra 63.953.674.2 Fable 5.1 55.654.656.6 Opus 5 52.251.053.3 Gemini 3.1 Pro 45.845.945.8 Grok 4.6 43.539.747.3 Kimi K3 40.541.140.0 Muse Spark 1.1 37.249.424.9 GLM 5.3 35.639.132.0 Grok 4.5 32.339.325.4 DeepSeek 4 Pro 26.432.420.3 Vision Capabilities Index Size: StandardMini Model| GPT-6 Astra 83.082.392.793.177.469.7 Fable 5.1 68.970.384.577.967.744.2 Opus 5 68.070.884.375.865.143.9 Gemini 3.1 Pro 62.174.284.166.153.632.4 Kimi K3…
saved by
related reading
- AI in 2025: gestalt — LessWronglesswrong.com
- GLM-5.2 Risk Evaluation Report – SaferAIsafer-ai.org
- Field map – AISafety.comaisafety.com
- AI Model & API Providers Analysis | Artificial Analysisartificialanalysis.ai
- Robocurve — Real-world evaluations of physical AIrobocurve.org
- Claude Opus 4.5: Model Card, Alignment and Safetythezvi.substack.com
- [2602.12316] GT-HarmBench: Benchmarking AI Safety Risks Through the Lens of Game Theoryarxiv.org
- GPT-6 Astra System Carddeploymentsafety.openai.com
- Conceptual Reasoning Indexconceptualreasoning.ai
- Field map – AISafety.comaisafety.com
- BenchmarkList: Track the Frontier of AI Capabilitiesbenchmarklist.com
- Intelligenceintelligence.ai