# OpenAI Aims for Top Spot with New ‘Open’ AI Model

## OpenAI Aims for Top Spot with New ‘Open’ AI Model

OpenAI is planning to release its first truly “open” language model since GPT-2 later this year, and new details are emerging about the company’s ambitions for the project. According to sources within the AI developer community, OpenAI is targeting an early summer release for the model, spearheaded by VP of Research Aidan Clark. The company’s goal is nothing short of creating the best-in-class open reasoning model, surpassing existing open alternatives.

This new model will reportedly be a “text in, text out” system, designed to run on high-end consumer hardware. Developers may also have the option to toggle the model’s “reasoning” capabilities on and off, similar to features seen in models recently released by Anthropic. While reasoning enhances accuracy, it can also increase latency, offering users a trade-off between speed and precision.

Furthermore, OpenAI is considering a highly permissive license for the model, aiming to avoid the “onerous requirements” that have drawn criticism toward other open models like Llama and Google’s Gemma. This move signals a potential shift in OpenAI’s approach to open source, especially as the company faces increasing competition from rivals like Chinese AI lab DeepSeek, which have successfully leveraged open strategies to build large user bases. Meta’s Llama models, for example, have amassed over 1 billion downloads.

OpenAI CEO Sam Altman has publicly acknowledged that the company may have been on the “wrong side of history” regarding open source, suggesting a renewed commitment to a more open approach. This shift includes plans for thorough red-teaming and safety evaluations of the upcoming model, with the release of a detailed model card outlining internal and external benchmarking and safety testing results.

Altman emphasized the importance of safety, stating that the model will be evaluated according to OpenAI’s preparedness framework, with extra attention given to its post-release modifiability. This focus on safety comes after criticism surrounding the reported rushing of safety testing on recent models and the failure to release model cards, underscoring OpenAI’s intent to prioritize responsible AI development with its new open model. The company was contacted for comment but has yet to respond.