In a surprising development, Anthropic has now overtaken OpenAI in enterprise usage of large language models (LLMs), according to recent usage metrics published by several industry sources. Claude, Anthropic’s flagship AI model, now commands roughly 32% of enterprise LLM use, compared to OpenAI’s GPT models, which have dropped to around 28–30%.
The shift is driven by Claude’s focus on safety, reliability, and clear boundaries in sensitive environments like healthcare, finance, and legal tech. Enterprise customers cite fewer hallucinations, better explainability, and stronger compliance features as key advantages. Notably, Claude 3.5’s release earlier this year included new safety frameworks that appeal to risk-averse sectors.
OpenAI, while still a leader in consumer AI via ChatGPT, is facing increased scrutiny over data governance and content generation practices. Meanwhile, Microsoft has begun integrating Claude into Azure alongside OpenAI models, offering more choice to enterprise clients.
This change underscores a broader trend in the AI industry: enterprises are prioritising model reliability and responsible AI practices over raw capability. As regulations tighten, companies are choosing models that align with their internal policies and legal requirements.
While OpenAI remains dominant in public discourse, especially with GPT-4o, Anthropic’s rise signals a maturing market where performance, safety, and governance take centre stage. Analysts expect continued fragmentation in the enterprise AI space as new players like Mistral and Cohere also gain ground.