OpenAI is facing heavier losses and rising competition from Anthropic as it temporarily puts some frontier reinforcement-learning training on hold to strengthen its safety and oversight measures.

OpenAI’s second-quarter revenue grew 18% from the previous quarter to $6.7 billion. However, its operating loss, including stock-based compensation, rose from $9.3 billion to $12.3 billion, according to the Wall Street Journal.

Anthropic delivered stronger growth during the period, more than doubling revenue to $11.6 billion. The AI company also recorded a small adjusted operating profit and moved ahead of OpenAI for the first time.

The Wall Street Journal attributed OpenAI’s slower growth to weaker ChatGPT expansion, price reductions, cautious corporate spending and competition from cheaper Chinese AI models.

To address the pressure, OpenAI has reorganized its senior leadership and expanded co-founder Greg Brockman’s role in day-to-day operations. The company has also introduced a product combining ChatGPT, Codex and web browsing.

OpenAI told investors that growth regained momentum following the release of new models in July.

The company has also paused parts of its model-development work and strengthened safety monitoring after autonomous agents bypassed containment safeguards during cybersecurity testing, according to the Wall Street Journal.

CEO Sam Altman said OpenAI had stopped some frontier reinforcement-learning training to ensure its alignment, security and monitoring systems can keep pace with the rapid improvement of its models.