The Archive

OpenAI and Microsoft sign new MOU reinforcing partnership on AI safety and innovation.

OpenAI restructures nonprofit governance with equity grants enabling $100B+ in resources for safe AI development.

SafetyKit product leverages GPT-5 for content moderation and compliance enforcement with improved accuracy over legacy systems.

OpenAI opens $50M People-First AI Fund applications for US nonprofits focusing on education, community innovation, and economic opportunity.

OpenAI research identifies mechanistic causes of language model hallucinations and proposes improved evaluation methods for reliability.

OpenAI announces GPT-5 bio bug bounty program with up to $25K rewards for identifying safety vulnerabilities via universal jailbreaks.

OpenAI partners with Greek Government to deploy ChatGPT Edu in secondary schools and support local AI ecosystem development.

OpenAI launches Jobs Platform and AI Certifications to increase worker access to training and employment opportunities.

Vijaye Raji appointed CTO of Applications at OpenAI following Statsig acquisition; reports to Fidji Simo.

OpenAI releases gpt-realtime speech-to-speech model with Realtime API updates including MCP support, image input, and SIP calling.

OpenAI announces $50M People-First AI Fund for U.S. nonprofits in education, healthcare, and research; applications Sept–Oct 2025.

OpenAI's survey of 1,000+ people worldwide reveals alignment between public values and Model Spec; informs AI behavior defaults.

OpenAI and Anthropic publish joint safety evaluation results on misalignment, hallucinations, jailbreaking, and instruction following.

OpenAI addresses safety protocols for users in mental/emotional distress, acknowledges system limits, and outlines refinement efforts.

OpenAI announces Learning Accelerator program (details not provided).

OpenAI's GPT-4b micro aided Retro Bio in protein engineering for stem cell therapy, demonstrating domain-specific model utility in life sciences.

OpenAI urges California Governor Newsom to harmonize state AI regulation with national standards, positioning federal leadership.

Basis built AI agents using o3, o3-Pro, GPT-4.1, and GPT-5 delivering 30% time savings for accounting firms.

OpenAI releases GPT-5 in API with advanced reasoning, developer controls, and improved coding task performance.

OpenAI demonstrates GPT-5 applied to medical research workflows and use cases.

OpenAI announces GPT-5, achieving state-of-the-art performance in coding, math, writing, vision, and reasoning across benchmarks.

OpenAI introduces output-centric safety training in GPT-5 that replaces hard refusals with nuanced safe-completions for dual-use prompts.

OpenAI and GSA launch ChatGPT Enterprise across U.S. federal executive branch at minimal cost.

OpenAI research quantifies worst-case frontier risks of open-weight LLMs via malicious fine-tuning attacks on biology and cybersecurity domains.

OpenAI releases open-weights models to broaden AI accessibility and adoption globally.

OpenAI ships gpt-oss-120b and gpt-oss-20b open-weight reasoning models under Apache 2.0.

OpenAI releases gpt-oss-120b and gpt-oss-20b open-weight models under Apache 2.0, optimized for reasoning and consumer hardware deployment.

OpenAI launches Stargate Norway, first European AI data center under OpenAI for Countries program.