OpenAI is gearing up to release its first open language model since GPT-2, led by VP of Research Aidan Clark. Slated for an early summer launch, the model aims to outperform existing open reasoning models. OpenAI is considering a permissive license to avoid heavy restrictions faced by competitors like Llama and Gemma. With rivals adopting open strategies, OpenAI's model will be designed for consumer hardware and may feature adjustable reasoning capabilities. The release will include rigorous safety evaluations and a comprehensive model card, addressing previous criticisms of rushed safety testing.