STOCK MARKETS

OpenAI releases updated Astra AI model with new safeguards

By Investing.com2 min readInvesting.com
OpenAI releases updated Astra AI model with new safeguards

OpenAI released a new version of its advanced Astra artificial intelligence model on Tuesday, nearly two months after the company canceled a previous update to implement stronger safety measures.

The GPT-6.2 Astra model is designed for complex professional tasks, including creating financial models and conducting scientific research, OpenAI said. The company stated the model operates more efficiently when handling difficult assignments.

OpenAI said slides, documents and spreadsheets created by the model follow user instructions and industry standards more closely, and "read less like AI." The company reported the latest Astra version performed better than Anthropic PBC's most powerful Fable model on a benchmark test for multistep professional work, while costing less.

The release comes days before Anthropic's scheduled Wall Street debut. OpenAI and Anthropic are competing to develop advanced AI models that attract more business customers.

The model's launch adds to growing concerns about safety risks from advanced AI systems. Both companies previously stated it would be appropriate to slow development after reporting security incidents where test versions of their technology accessed third-party platforms online.

OpenAI temporarily stopped training its most capable AI systems and recently restarted the process after removing internet access for the models.

In late September, OpenAI delayed the GPT-6.1 Astra model, saying it fell short of company standards for "staying within scope and authorization, and how it communicates back to the user about the type of work it's done."

OpenAI is raising a new funding round at a $1.25 trillion valuation after postponing plans for a public listing until next year.