The Reign of ChatGPT in LLM Usage - A Deep Dive

86% of all LLM usage is driven by ChatGPT - Help Net Security

ChatGPT remains the most widely used LLM among New Relic customers, comprising over 86% of all tokens processed. Developers and enterprises are transitioning towards OpenAI's latest models, such as GPT-4o and GPT-4o mini, despite the availability of more cost-effective alternatives.

The Dominance of ChatGPT

Speed, reliability, and enterprise readiness are the key factors driving ChatGPT's dominance. Many organizations opt for deploying established models rather than investing in training or fine-tuning their own. However, continuous monitoring remains crucial, even for well-established models, to track usage, benchmark performance, and detect anomalies for cost management and reliability maintenance.

Shift Towards OpenAI's Updates

Enterprises are quick to adopt OpenAI's updates. New Relic users have transitioned from ChatGPT-3.5 Turbo to the newer ChatGPT-4.1 mini since April, prioritizing enhanced performance and features over cost considerations.

Rising Model Diversity

While ChatGPT continues to lead the race, data indicates a growing model diversity. Meta's Llama has emerged as the second most used model, with a 92% increase in the number of unique models utilized across AI applications in Q1 2025. Developers are exploring various alternatives, including open-source and specialized solutions, albeit on a smaller scale.

Importance of Monitoring in AI Adoption

AI adoption is fueling the demand for monitoring. As AI operations become mainstream, observability tools need to encompass the entire AI stack beyond LLMs. In the absence of proper monitoring, companies face risks such as performance degradation, security vulnerabilities, and escalating costs, potentially hindering AI innovation and eroding confidence in scaling AI.

The Role of Programming Languages

Python continues to lead in terms of momentum, support, and tooling, dominating AI applications, with customer adoption surging by almost 45% since the last quarter. Following Python, Node.js ranks high in both the scale of requests and customer adoption. Notably, Java usage has witnessed a rapid 34% increase since the previous quarter, indicating a forthcoming surge in production-grade, Java-based LLM applications from major enterprises.