OpenAI surprises top researchers and engineers with multimillion-dollar bonuses ahead of GPT-5 launch

Just before the launch of GPT-5, OpenAI CEO Sam Altman surprised employees with a surprise announcement in the company’s Slack channel. In a message shared with reporters, Altman revealed that the company would be awarding a “special one-time” bonus to researchers and software engineers in select teams, including applied engineering, scaling, and safety. The bonuses, which will be paid out quarterly over the next two years, are expected to range from hundreds of thousands of dollars for engineers to mid-to-single-digit millions for top-tier researchers. Employees can choose to receive the payouts in cash, OpenAI stock, or a combination of both. Approximately 1,000 employees—about one-third of OpenAI’s full-time workforce—are eligible. This marks the first time OpenAI has issued such a broad, high-value bonus to technical staff, underscoring the company’s urgency to retain talent amid intense competition. Altman emphasized that while compensation will continue to rise as the company succeeds, this particular award was being shared transparently because it was a new initiative. He also thanked employees for their contributions to advancing artificial general intelligence. The move comes amid a wave of talent poaching. OpenAI has recently lost several key researchers to Meta, including Shengjia Zhao, one of ChatGPT’s original creators, who was named chief scientist at Meta. Mark Chen, OpenAI’s chief research officer, has privately likened Meta’s recruitment efforts to a “home invasion,” while Altman has told staff that “missionaries will beat mercenaries” in the battle for AI’s future. Elon Musk’s xAI has also been aggressively targeting OpenAI’s top engineers, and ex-CTO Mira Murati recently recruited several technical leaders to join her new venture, Thinking Machines. The bonus program may create some tension among employees who were excluded, particularly those who worked tirelessly on the GPT-5 launch. Still, morale appears high following the event, with Altman joining staff for a celebratory happy hour at Splash, a sports bar near OpenAI’s San Francisco headquarters. Meanwhile, Anthropic is facing setbacks. Despite earlier enthusiasm, several companies, including Cursor, have now switched to using GPT-5 as their default model, replacing Anthropic’s Claude. This shift is significant given Cursor’s role as a major revenue source for Anthropic. In response, Anthropic rushed out an update to improve Claude’s coding abilities. Early access to GPT-5 was limited to a select group of independent writers and AI influencers under strict non-disclosure agreements—none of whom are from traditional media outlets like The Verge. The chosen few, including Simon Willison, Ethan Mollick, and teams from Every and Latent Space, have shared early impressions, with most praising the model’s advanced reasoning and coding skills. On the valuation front, Bloomberg reported that OpenAI may be valued at $500 billion in an upcoming tender offer, far exceeding the $300 billion valuation from its last funding round. Altman has privately told employees they could sell their vested shares at significantly higher prices than the $274 per share previously offered.