Lab system cards and frameworks

System cards, model cards, safety frameworks and policies from the labs that train frontier models.

60 of 329 documents, newest first

System card · Nov 2026GPT-6 Cyber System CardOpenAI · 140 paragraphs · Read at source System card · Nov 2026ChatGPT Images 2.6 System CardOpenAI · 58 paragraphs · Read at source Model card · Nov 2026Gemini Omni 1.2 Flash Model CardGoogle DeepMind · 48 paragraphs · Read at source System card · Oct 2026GPT-Rosalind-6 System Card AddendumOpenAI · 107 paragraphs · Read at source Framework · Oct 2026OpenAI’s Raising Concerns PolicyOpenAI · 8 paragraphs · Read at source System card · Oct 2026Addendum to GPT-6 Luna System Card: October 29, 2026 UpdateOpenAI · 100 paragraphs · Read at source Article · Oct 2026An updated Model Spec for more capable and beneficial AIOpenAI · 21 paragraphs · Read at source System card · Oct 2026Addendum to GPT-6 Astra System Card: GPT-6.1 TerraOpenAI · 255 paragraphs · Read at source Article · Oct 2026Broad access through consistent standardsOpenAI · 14 paragraphs · Read at source Framework · Oct 2026Frontier Safety Roadmap: archive of previous goalsAnthropic · 100 paragraphs · Read at source Framework · Oct 2026Frontier Safety RoadmapAnthropic · 61 paragraphs · Read at source System card · Oct 2026GPT-Live-1.1 system cardOpenAI · 51 paragraphs · Read at source Article · Oct 2026Distillation defensesAnthropic · 31 paragraphs · Read at source Article · Oct 2026Disrupting AI-enabled “false front” operationsOpenAI · 67 paragraphs · Read at source System card · Oct 2026System Card: Claude Haiku 5.5Anthropic · 144 pages · Read at source System card · Oct 2026GPT-6 Sol and GPT-6 Luna: October 2026 updateOpenAI · 152 paragraphs · Read at source Article · Oct 2026Transparency HubAnthropic · 793 paragraphs · Read at source Model card · Oct 2026Gemini 4 Argon Model CardGoogle DeepMind · 14 pages · Read at source Model card · Oct 2026Nano Banana 2.1 Model CardGoogle DeepMind · 8 pages · Read at source Article · Sep 2026Disrupting a coordinated model-distillation campaignOpenAI · 17 paragraphs · Read at source System card · Sep 2026Addendum to GPT-6 Astra System Card: GPT-6.1 SolOpenAI · 262 paragraphs · Read at source System card · Sep 2026System Card: Claude Sonnet 5.5Anthropic · 148 pages · Read at source System card · Sep 2026System Card: Claude Opus 5.5Anthropic · 230 pages · Read at source Article · Sep 2026Building standards for the next phase of AIOpenAI · 34 paragraphs · Read at source Article · Sep 2026Introducing the Australian Youth Safety BlueprintOpenAI · 5 paragraphs · Read at source Framework · Sep 2026Amazon's Frontier Model Safety Framework (September 2026)Amazon · 11 pages · Read at source Framework · Sep 2026Our framework for reporting model misalignmentOpenAI · 33 paragraphs · Read at source Article · Sep 2026The AI policy window is open. We need to act.OpenAI · 44 paragraphs · Read at source System card · Sep 2026ChatGPT Images 2.5 System CardOpenAI · 53 paragraphs · Read at source System card · Sep 2026GPT-6 Astra System CardOpenAI · 1,010 paragraphs · Read at source System card · Sep 2026Safety overview: GPT-6 AstraOpenAI · 12 paragraphs · Read at source System card · Sep 2026System Card: Claude Fable 5.1 & Claude Mythos 5.1Anthropic · 212 pages · Read at source Article · Sep 2026Enterprise frontier safeguardsAnthropic · 28 paragraphs · Read at source Article · Sep 2026Detecting and countering misuse of AI: September 2026 (threat intelligence report)Anthropic · 154 pages · Read at source Model card · Sep 2026Gemini 3.8 Audio (Live, Live Extended Thinking, Flash TTS, Flash-Lite TTS) Model CardGoogle DeepMind · 7 pages · Read at source Model card · Sep 2026Gemini 3.8 Flash Model CardGoogle DeepMind · 8 pages · Read at source Article · Aug 2026OpenAI – Hugging Face Incident Technical ReportOpenAI · 38 pages · Read at source Article · Aug 2026Disrupting a new covert influence campaign from RussiaOpenAI · 22 paragraphs · Read at source Article · Aug 2026Strengthening democratic oversight in national securityOpenAI · 20 paragraphs · Read at source Article · Aug 2026New policy ideas for the Intelligence AgeOpenAI · 25 paragraphs · Read at source Article · Aug 2026Claude text watermarkAnthropic · 54 paragraphs · Read at source Article · Aug 2026Redacted Risk Report, August 2026Anthropic · 186 pages · Read at source Framework · Aug 2026Anthropic’s Responsible Scaling PolicyAnthropic · 98 paragraphs · Read at source Article · Aug 2026Improving Fable 5's biology safeguardsAnthropic · 19 paragraphs · Read at source System card · Aug 2026GPT-5.6 — August UpdatesOpenAI · 187 paragraphs · Read at source Model card · Aug 2026Gemini 3.5 Audio (Live Translate, Transcribe, Transcribe Live) Model CardGoogle DeepMind · 7 pages · Read at source Model card · Aug 2026Gemini 3.7 Flash Model CardGoogle DeepMind · 9 pages · Read at source Article · Jul 2026Disrupting a Cambodia-based criminal scam operationOpenAI · 25 paragraphs · Read at source Evaluation report · Jul 2026Investigating incidents in cybersecurity evaluationsAnthropic · 53 paragraphs · Read at source Evaluation report · Jul 2026Gemini Robotics 2 Safety ReportGoogle DeepMind · 18 pages · Read at source Article · Jul 2026Position on open-weights modelsAnthropic · 12 paragraphs · Read at source System card · Jul 2026System Card: Claude Opus 5Anthropic · 194 pages · Read at source Article · Jul 2026Transparency Hub: System Trust and ReportingAnthropic · 53 paragraphs · Read at source Article · Jul 2026Hard questionsAnthropic · 15 paragraphs · Read at source System card · Jul 2026GPT-5.6 System CardOpenAI · 492 paragraphs · Read at source Article · Jul 2026Our approach to government and national security partnershipsOpenAI · 9 paragraphs · Read at source System card · Jul 2026GPT-Live System CardOpenAI · 45 paragraphs · Read at source Model card · Jul 2026Gemini 3.5 Flash-Lite Model CardGoogle DeepMind · 7 pages · Read at source Model card · Jul 2026Gemini 3.6 Flash Model CardGoogle DeepMind · 7 pages · Read at source Model card · Jul 2026Gemini Robotics ER 2 Model CardGoogle DeepMind · 5 pages · Read at source