OpenAI partners with Hacktron for vulnerability research

OpenAI has partnered with security firm Hacktron to jointly test vulnerabilities in its systems, particularly focusing on third-party service interactions. The deal grants Hacktron expanded model access through the Daybreak Red program as part of a broader security strategy.

By Nadia Vale

November 6, 2026 at 6:34 PM EST1 min read

Key takeaways (3)

1.OpenAI and Hacktron will jointly test AI agent interactions with third-party services.
2.Hacktron gained enhanced access to OpenAI's frontier models through Daybreak Red.
3.The announcement includes no specific vulnerabilities or test results.

Key takeaways (3)

1.OpenAI and Hacktron will jointly test AI agent interactions with third-party services.
2.Hacktron gained enhanced access to OpenAI's frontier models through Daybreak Red.
3.The announcement includes no specific vulnerabilities or test results.

OpenAI announced a partnership with security firm Hacktron on November 6, 2026, to identify vulnerabilities in its systems through joint testing of AI agent interactions.

The collaboration expands Hacktron's access to OpenAI's frontier models via the Daybreak Red program. The two organizations will jointly select high-priority areas for testing, with a specific focus on how OpenAI's systems interact with third-party services. This initiative follows separate announcements earlier this week regarding a security sprint with Cathedral and increased bug bounty rewards for independent researchers.

Under the terms of the partnership, OpenAI is granting Hacktron enhanced access to its most advanced models. The company stated that engineers and Hacktron's security researchers will choose high-priority areas to test together. The primary objective is to evaluate the security implications of AI agents connecting various services in new ways.

"As AI agents connect services in new ways, that combination of skills becomes even more valuable. Working alongside Hacktron will help us test those connections more deeply and use what we learn to strengthen our defenses," OpenAI said in its announcement [Partnering with Hacktron to find vulnerabilities in OpenAI's systems](https://openai.com/index/partnering-with-hacktron/).

Hacktron builds AI agents designed for code review, penetration testing, and red teaming. According to OpenAI, the firm's previous research has uncovered vulnerabilities in applications, cloud infrastructure, and identity systems. The company noted it became familiar with Hacktron through its Security Bug Bounty program, citing the impact of vulnerabilities previously discovered and disclosed by the firm.

While the partnership focuses on third-party service interactions, it does not detail the specific technical methodologies Hacktron will employ or the exact scope of the Daybreak Red access. The announcement directs readers to the Security Bug Bounty program 1 for information on which systems are currently in scope for general submissions.

NVIDIA says Vera CPU runs agents 2x faster than most efficient x86 chip

The announcement comes as businesses place more AI agents into production to search databases, write code and take action.

Adrian Cole · about 24 hours ago

Nanbeige releases open-weights 33B model for agentic tasks

Kanzhun’s Nanbeige lab released a 33B open-weights model under Apache 2.0, matching Qwen3.8 27B on an AI index while showing stronger terminal use and weaker knowledge benchmarks. The model is described as a teacher for future distillation into smaller agents.

Adrian Cole · 1 day ago

NVIDIA adds four open-weight Cosmos 3 Nano checkpoints for interactive simulation

The models target robotics and autonomous-driving researchers who need controllable, photorealistic environments to train and evaluate policies without real-world risk.

Mira Sethi · 1 day ago