OpenAI debuted its latest model GPT-5.5, boasting improved agentic coding, computer use, and early-stage research capabilities. Meanwhile, Nvidia has provided early access to its 10,000 employees for GPT-5.5 testing.

OpenAI has announced the release of GPT-5.5, an improvement over the existing model, emphasizing multi-step work and agentic coding [doc10]. The new version can plan, use tools, and verify its output more effectively with less assistance from humans. GPT-5.5 offers faster assistance for complex problems, while the Pro version is aimed at researchers who need accuracy over speed.

In related news, Nvidia granted its 10,000 employees early access to GPT-5.5 through Codex, positioning the company at the forefront of the model's deployment. GPT-5.5 has also been pitched as more token-efficient for Codex tasks.

DeepSeek, a Chinese AI company, has released its latest V4 AI model, boasting 'world-class' reasoning capabilities, with a context length of up to 1 million [doc12] competing with OpenAI's GPT-5.5. This move could pose competition to OpenAI's latest model.

However, the introduction of GPT-5.5 has led to market analysis indicating that Nvidia's stock may face challenges, especially with regards to its position in the AI development.

Microsoft's investment in OpenAI has been noted as a significant factor that may have contributed to this recent breakthrough in AI.

Nvidia Teams Up with OpenAI for GPT-5.5 and OpenAI Debuts GPT-5.5