Artificial intelligence (AI) company OpenAI has released its latest model, GPT-5.3-Codex, which shows significant improvements over rival systems in coding benchmarks and testing results. However, the company is delaying full access to developers due to concerns about cybersecurity risks.
GPT-5.3-Codex can perform tasks like writing, testing, and reasoning about code with high accuracy, but this also raises serious concerns about its potential misuse for malicious activities. OpenAI has implemented additional safety measures, including a trusted-access program for vetted security professionals, to mitigate these risks.
The company is offering $10 million in API credits to developers working on cybersecurity-related projects, highlighting its commitment to responsible AI development. This move aims to balance the benefits of advanced AI coding capabilities with the need to protect against potential cyber threats.
While OpenAI’s model shows impressive performance, the company acknowledges that it has reached a critical point where it requires more stringent controls and safeguards to prevent misuse. By prioritizing cybersecurity, OpenAI is taking a proactive approach to ensuring that its technology is used for good, rather than harm.
Source: https://fortune.com/2026/02/05/openai-gpt-5-3-codex-warns-unprecedented-cybersecurity-risks