Arize AI offers ML and LLM observability with tracing, evals, and drift detection. Compare its free Phoenix open source tool and AX Pro pricing for 2026.
Arize AI is an AI and machine learning observability company that helps engineering teams monitor, evaluate, and debug the performance of models and AI applications running in production. It was founded in January 2020 by Aparna Dhinakaran and Jason Lopatecki and is headquartered in Berkeley, California.
The company began with a focus on traditional ML monitoring, including drift detection and model performance dashboards, and has since expanded heavily into large language model and AI agent observability as generative AI adoption has grown across enterprises.
Arize has raised approximately 131 million dollars across four funding rounds, including a 70 million dollar Series C led by Adams Street Partners, with participation from investors such as Microsoft's M12, Datadog, and PagerDuty.
Arize AX, the company's enterprise platform, provides session-level and span-level tracing for LLM and AI agent workflows, letting teams see exactly how a request flowed through a pipeline of model calls, tools, and retrieval steps.
The platform includes LLM-as-a-judge evaluation capabilities for automatically scoring model outputs, real-time alerting integrated with tools like PagerDuty and Slack, drift and data quality monitoring, and an AI debugging assistant called Alyx.
Arize also maintains Phoenix, a free and open-source observability and tracing library that has surpassed two million monthly downloads, giving teams a no-cost, self-hosted entry point into LLM observability before adopting the full commercial platform.
Arize follows a freemium pricing model with four tiers. Phoenix is fully open source and free to self-host with no usage caps. AX Free is a free managed cloud plan offering 25,000 spans per month, 1 GB of ingestion, and 15-day retention.
AX Pro costs 50 dollars per month for 50,000 spans and 10 GB of ingestion with 30-day retention, with additional usage billed per million spans or per gigabyte of storage. AX Enterprise offers custom SaaS or self-hosted pricing with extended retention, dedicated support, compliance certifications, and SSO, with typical enterprise contracts reported around 60,000 dollars per year.
Arize AI provides observability and evaluation tools that help teams monitor, evaluate, and debug the performance of ML models and LLM/AI agent applications in production.
Yes, Arize offers a fully free and open-source tool called Phoenix, plus a free managed AX Free tier with 25,000 spans per month.
Arize AI was founded in January 2020 by Aparna Dhinakaran and Jason Lopatecki.
Arize AX Pro costs $50 per month for 50,000 spans and 10 GB of ingestion, with additional usage billed separately.
Phoenix is Arize's free, open-source AI observability and tracing tool, which can be self-hosted with no usage caps.
Arize is used by ML engineering teams, AI platform teams, and enterprises operating LLM and AI agent applications in production.
Arize has raised approximately $131 million across four funding rounds, including a $70 million Series C.
Alyx is Arize AX's built-in AI debugging assistant that helps engineers diagnose production issues faster.