4 links
tagged with all of: openai + gpt-4.1
Click any tag below to further narrow down your results
Links
OpenAI has released its new AI model, GPT-4.1, which reportedly outperforms some previous models in programming benchmarks, but it has not accompanied this release with a safety report, diverging from industry norms. The lack of a system card has raised concerns among safety researchers, particularly as AI labs are criticized for lowering their reporting standards. Transparency in AI safety assessments remains a voluntary commitment by companies like OpenAI, despite their previous pledges for accountability.
OpenAI has launched GPT-4.1, a new family of generative multimodal AI models that can handle 1 million tokens of context. This release includes three versions: GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano, with the latter being described as the smallest and fastest model. Improvements from GPT-4.1 will be integrated into ChatGPT, while the previous model, GPT-4, will be phased out by April 30th.
The article discusses the safety report for OpenAI's GPT-4.1, highlighting the results from the SPLXAI test that evaluated the model's performance in various scenarios. It emphasizes the advancements made in ensuring the model's outputs are safer and more aligned with user expectations while addressing potential risks and areas for improvement.
OpenAI has announced new models under the GPT-4.1 banner that focus specifically on coding capabilities. These updates aim to enhance programming tasks, making it easier for developers to integrate AI into their workflows and improve coding efficiency.