LLM observability & evaluation · Arize AI

Arize Phoenix

Open-source, local-first tracing and evaluation library for LLM and agent applications, with an optional path to Arize's commercial platform.

Visit Arize Phoenix Alternatives

Arize Phoenix is an open-source library for tracing, evaluating, and iterating on LLM and agent applications, designed to run locally (in a notebook or as a local server) before any data leaves a developer's machine. It supports OpenTelemetry-based tracing of chains and agent steps, built-in LLM-as-judge evaluators, prompt iteration and comparison, and dataset-based experiments, aimed at the debugging and evaluation phase of building an application. Phoenix itself has no pricing; it is free and self-hosted. Arize AX is the company's separate commercial, managed observability platform for production monitoring, which shares evaluation concepts with Phoenix but adds hosted infrastructure, drift/data-quality signal detection, and enterprise security controls, billed by monthly trace-span volume and data ingestion.

At a glance

Vendor Arize AI
Pricing model Open source + paid options
Free tier Yes
Deployment Cloud, Self-hosted
Open source Yes (Apache-2.0)
Best for Developers who want a free, local, open-source tracing and eval library during development before committing to a hosted platform.

Pricing

Phoenix itself is free, open-source, and local-first with no hosted fees; Arize AX, the company's separate commercial production-monitoring platform, has its own free, Pro, and Enterprise plans metered by trace spans and data ingestion.

Plan Price Notes
Phoenix (open source) Free Self-hosted or local; no usage limits from Arize
Arize AX Free $0 25k trace spans/month, 1 GB ingestion/month, 15-day retention
Arize AX Pro $50/month 50k trace spans/month, 10 GB ingestion/month, 30-day retention
Arize AX Enterprise Custom Unlimited spans/ingestion, self-hosted or SaaS, SSO, audit logs, SOC 2 Type II

Prices read from the vendor's own page on September 21, 2026. Vendors change prices; check the source before you budget.

Features

  • OpenTelemetry-based tracing for LLM and agent applications
  • Built-in LLM-as-judge and heuristic evaluators
  • Local-first, notebook-friendly workflow
  • Prompt iteration and comparison tools
  • Dataset-based experiments
  • Optional upgrade path to Arize AX for hosted production monitoring

Integrations

Profile last reviewed September 21, 2026

Alternatives

Langfuse Open-source LLM tracing and evaluation platform, available self-hosted for free or as a metered hosted cloud service. OpenLLMetry Apache-2.0 open-source instrumentation library extending OpenTelemetry to LLM calls, usable with any OpenTelemetry-compatible backend. W&B Weave LLM tracing and evaluation toolkit from Weights & Biases, sharing billing and infrastructure with its ML experiment-tracking platform. Ragas Open-source Python framework of reference-free metrics purpose-built for evaluating retrieval-augmented generation (RAG) pipelines.

Arize Phoenix in the index now

Terms to know

Data observability Continuous automated monitoring of pipelines and datasets to catch freshness, volume and quality problems early. Data quality The degree to which data is accurate, complete, consistent and fit for the purpose it is being used for. Machine learning Building models that learn patterns from data to make predictions or decisions, rather than following hand-coded rules. Data ingestion Moving data from its source systems into a storage or processing system where it can be transformed and analyzed. LLM observability Monitoring a language model's live behavior in production, tracking its inputs, outputs, latency, cost, and failure patterns.

Related guides

How to choose an LLM observability tool Pick an LLM tool by whether you need tracing, scoring or a gateway first — most teams eventually need all three, rarely from one vendor. How to evaluate LLM applications Offline evals, LLM-as-judge, human review and production monitoring, aimed at the whole pipeline rather than the model alone.