OpenAI has announced the forthcoming release of its Astra model, which the company claims is the first large language model to meet its “critical cybersecurity threshold.” According to OpenAI, Astra demonstrates the capability to identify and exploit unknown security vulnerabilities in computer systems autonomously, without human guidance. The company reports that Astra achieved a perfect score on ExploitBench, a benchmark for assessing an LLM’s ability to exploit known system vulnerabilities, and further discovered and exploited two zero-day vulnerabilities in a modified test developed by OpenAI engineers.

In preparation for Astra’s release, OpenAI is implementing several safety measures. The company has begun enhancing the model’s harness to detect and prevent abusive behavior and jailbreaks. Additionally, OpenAI is identifying “accounts assessed as higher risk” and restricting the model’s responses to their prompts, although the specifics of this process are not detailed. The company also plans to deploy Astra with chain-of-thought monitoring to identify and halt any malicious activity. OpenAI intends to provide a limited preview of Astra to a select group of testers before its wider public release, though the criteria for selecting these testers remain undisclosed. The company acknowledges the difficulty in independently verifying its claims about the model’s safety and preparedness at this stage.

This is a summary. Read the original at techcrunch.com.