Back to vendors
A

Arize AI

Also known as: Arize AX, Phoenix

Visit site
Entry priceAX Pro from $50 a month, plus span and ingestion overage. AX Free and Phoenix open source are free, and Enterprise is custom.Full pricing detail

AI observability and evaluation platform that traces, evaluates, and monitors LLM applications and agents, with an open source core in Phoenix.

Arize AI is an observability, evaluation and improvement platform for AI agents, sold in two forms on the same open standards. Dynatrace has completed its acquisition of Arize, and Arize says it will keep supporting Phoenix and AX. Phoenix is the open-source project, run locally, in containers or in the customer's cloud, with tracing, evaluation, datasets, experiments, a playground, prompt management, a built-in assistant, a CLI and a remote MCP server. Arize AX is the managed platform, backed by the adb datastore, which syncs trace data to BigQuery, Databricks or Snowflake.

AX traces agents on OpenTelemetry and OpenInference, the semantic conventions Arize maintains, with path and graph views of each trajectory, cost and latency, sessions, dashboards and monitors. It runs online evals on production traces and offline evals on datasets, including trace, session and agent-as-a-judge evals, human annotation and labeling queues, and Prompt Learning optimization. Two built-in agents work on the customer's agents. Alyx is an AI engineering agent that runs evals and debugs issues, and Signal surfaces failure patterns from production telemetry, traces root causes and proposes review-ready fixes.

Arize reports SOC 2 Type II, ISO 27001, PCI DSS, HIPAA and GDPR certification, offers US, EU or CA data regions, and adds self-hosted deployment, Enterprise SSO and audit logs on Enterprise. AX Free and Phoenix are free, AX Pro is $50 a month, and AX Enterprise is custom. Arize reports a trillion spans processed and a billion evaluations a year.

Vendor details

Canonical URL

https://arize.com

Category

Agent infrastructure

Subcategory

Observability and evaluation

Funding status

Founded January 2020 by Jason Lopatecki (CEO) and Aparna Dhinakaran (Chief Product Officer) in Berkeley, California. Has raised about $131M across rounds through Series C, and acquired Velvet in 2025. Dynatrace has completed its acquisition of Arize, for $915 million, and Arize says it will keep supporting Phoenix and AX.

Company status

acquired

Use cases & customers

Primary use cases

agent observabilityLLM evaluationagent tracingprompt optimizationproduction monitoring

Target customers

developersenterprise

Deployment options

SaaSself-hostedon-prem

Integrations

OpenTelemetry and OpenInference native instrumentation with framework integrations for OpenAI, Anthropic, LangChain, LangGraph, CrewAI, and LlamaIndex. Python and JavaScript SDKs, trace export to backends like Jaeger, Prometheus, and Grafana, adb Data Fabric sync to BigQuery, Databricks, and Snowflake, and a CLI usable from Cursor and Claude Code.

In practice

Your agent works in demos but fails unpredictably in production. You instrument it with Arize, trace every prompt, tool call, and route, and find the exact step where it breaks.

You want to know whether a prompt change actually improved quality, not just feels better. You run LLM as a judge evaluations on curated datasets and on live traffic, and compare results before shipping.

Your agent's failures in production go unnoticed until customers complain. Online evals run on live traces as they arrive, custom monitors fire on production conditions, and Signal surfaces failure patterns and proposes fixes for your team to review.

Agentic Index coverage score

9.0 / 14 capabilities · 64%

Integrations & Tool Calling Partial

More than 40 integrations with models, frameworks and AI tools instrument the customer's application to send traces into Arize, code evaluators run in E2B, Daytona, Vercel or Modal sandboxes, and Data Fabric syncs trace data to warehouses. These move telemetry in and data out, and no connector lets an agent take authenticated actions in outside systems.

SourceArize, arize.com homepage and pricing and the Phoenix README on github.com/Arize-airead 2026-09-21

Workflow Orchestration Not documented

Agents that Arize observes, evaluates and improves run elsewhere, and Arize has no workflows that sequence, branch or retry an agent's steps, or mix deterministic nodes with agent steps.

SourceArize, arize.com pricing and homepageread 2026-09-21

Knowledge Grounding & RAG Not documented

RAG pipelines built elsewhere are traced and evaluated in Arize, but Arize offers no retrieval structure over the customer's own documents that grounds an agent's answers.

SourceArize, arize.com homepage and pricingread 2026-09-21

Human Oversight & Guardrails Partial

Signal generates review-ready fixes for a person to review instead of applying them, so changes Arize's agents propose pass through human review. There is no approval step, escalation rule or pause before an agent acts at run time, and no autonomy modes that treat low risk and high risk actions differently.

SourceArize, arize.com homepage and pricingread 2026-09-21

Security, Identity & Governance Full

Arize reports SOC 2 Type II, ISO 27001, PCI DSS, HIPAA and GDPR certification, with a Trust Center, and its plans include organization- and space-level RBAC, service accounts, Enterprise SSO and audit logs.

SourceArize, arize.com homepage and pricingread 2026-09-21

Observability & Auditability Full

Agents are traced in Arize AX on OpenTelemetry with path and graph visualizations of each agent trajectory, token, latency and cost tracking, session and multi-modal tracing, custom dashboards, metrics, monitors and views on traces, audit logs on Enterprise, retention by plan (15 days Free, 30 days Pro, custom Enterprise), and Data Fabric sync of trace data to BigQuery, Databricks or Snowflake.

SourceArize, arize.com pricing and homepageread 2026-09-21

Memory & State Persistence Not documented

Arize stores agent trajectories and sessions for tracing and evaluation, and there is no session, conversation or long-term memory that an agent reads back as context.

SourceArize, arize.com homepage and pricingread 2026-09-21

Deployment & Data Residency Full

Every Arize AX plan chooses a US, EU or CA data region, Enterprise adds self-hosted deployments, and open-source Phoenix runs locally, in containers with Docker and Helm, or in the customer's cloud.

SourceArize, arize.com pricing and the Phoenix README on github.com/Arize-airead 2026-09-21

Prebuilt Agents, Templates & Packs Full

Ready-made agents a customer switches on include Alyx, an AI engineering agent that runs evals, debugs issues and improves the customer's agents, and Signal, whose agents find production issues and propose fixes, each working without the other, alongside managed agents on Enterprise and packaged skills that teach coding agents to trace, evaluate and debug with Phoenix. Those are role-specific agents and ready-made workflows.

SourceArize, arize.com homepage and pricing and the Phoenix README on github.com/Arize-airead 2026-09-21

Triggers & Channel Coverage Full

Online evals run on traces and spans as they arrive, custom monitors fire on production conditions, and Signal's agents surface failure patterns from production telemetry, trace root causes and generate proposed fixes, so Arize's work starts from incoming telemetry with no person initiating each run.

SourceArize, arize.com homepage and pricingread 2026-09-21

Model Flexibility & Routing Full

The AI provider behind Arize's own agents and evaluators can be customized, with the customer's own key alongside Arize-managed models on Enterprise, and the Phoenix playground compares models and replays traced LLM calls against them. Customers control the models Arize's own AI features run on, though bringing their own key requires the Enterprise tier.

SourceArize, arize.com pricing and the Phoenix README on github.com/Arize-airead 2026-09-21

APIs, SDKs & MCP Extensibility Full

Arize publishes Python and TypeScript client, OpenTelemetry and evals packages, a GraphQL API, a CLI that fetches traces, datasets and experiments, and a remote MCP server at each Phoenix instance's /mcp endpoint that MCP clients use to query traces, datasets and experiments, all on the OpenInference and OpenTelemetry standards it maintains. That is a stable API and SDK for Arize's own platform, with MCP served.

SourceArize, the Phoenix README on github.com/Arize-ai and arize.com homepageread 2026-09-21

Testing, Debugging & Optimization Full

Testing covers online evals on traces and spans and offline evals on datasets and experiments, with trace evals of agent trajectories, session evals of multi-turn conversations, agent-as-a-judge and custom code evaluators, evaluator alignment, human annotations and labeling queues, agent experimentation in the playground, and Prompt Learning optimization. Phoenix adds open-source evals, datasets and experiments. That tests the customer's agent against datasets before production and scores quality over time.

SourceArize, arize.com pricing and the Phoenix README on github.com/Arize-airead 2026-09-21

Browser & Computer Use Not documented

Arize observes, evaluates and improves agents, and its agents have no browser, desktop or computer control.

SourceArize, arize.com homepageread 2026-09-21

The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Product updates

·Agent capabilityVerified

Eval Health is a new managed agent in Arize AX that reviews the evals on a tracing project. It reads recent spans, tasks and evaluators, then files an issue for each coverage gap or broken eval with a proposed fix: create, repair, delete or optimize an eval, such as turning an LLM judge into a code eval to cut cost. Nothing changes until the user accepts the proposal.

Bears on: Observability / auditability

View source
·Agent capabilityVerified

Alyx, the agent built into Arize AX, can now create, edit, explain and run Decision Judge evaluators from chat. It reuses a matching evaluator or proposes a new one, sets up a project task, checks the column mappings and runs it over historical LLM spans, and nothing saves until the user accepts it.

Bears on: Observability / auditability

View source
·Observability / auditabilityVerified

Phoenix v20.18.0 shows evaluation results directly in the trace tree, so eval scores and labels sit beside each step of an agent or LLM run.

Bears on: Observability / auditability

View source
View all 8 Arize AI updates →Tracked since Jul 2026 · Verified from public vendor sources

Pricing

AX Pro from $50 a month, plus span and ingestion overage. AX Free and Phoenix open source are free, and Enterprise is custom.

A flat monthly plan with usage overage on trace spans and on ingestion, billed per GB.

Free tier

Included quota

AX Free includes 25,000 trace spans and 1 GB ingestion per month with 15 day retention. AX Pro includes 50,000 spans and 10 GB per month with 30 day retention. Enterprise allowances are custom.

What is public

Arize publishes full self serve pricing. AX Free and Phoenix open source are free, AX Pro is $50 per month, and AX Enterprise is custom. Span and ingestion allowances and overage rates are listed per tier.

Billing mechanics

Tiers meter trace spans and data ingestion per month with fixed retention windows. AX Pro is a flat $50 monthly fee that includes 50,000 spans and 10 GB. Beyond that, spans bill at $0.0008 each and ingestion at $3 per GB. Enterprise replaces fixed limits with negotiated allowances.

Cost watchouts

Span and ingestion overage on AX Pro can grow with high volume agents. The gap between $50 Pro and custom Enterprise is large with no middle tier.

Variable cost rationale

Arize meters trace spans and data ingestion. The $50 Pro plan includes 50k spans and 10 GB a month, and agents that send more accrue per span and per GB overage, so the bill grows with agent traffic and trace size. The flat fee and included allowances set the floor, and no overage cap is published. Heavy scale moves teams to a custom Enterprise contract with negotiated allowances, and Phoenix open source can be self hosted with no Arize fee.

Additional watchouts

Offline evaluation on data older than the tier retention window (15 or 30 days) is not available until Enterprise. Compliance certifications and enterprise SSO are Enterprise only.

Overage / add-ons

On AX Pro, additional trace spans are $0.0008 each and additional ingestion is $3 per GB. Enterprise overage is negotiated.

Sales call required

Mixed (some tiers require a call)

Free / trial

AX Free with 25k spans and 1 GB a month, 15 day retention and 10 Signal issues a month, plus free Phoenix open source

Lowest paid plan

AX Pro at $50 per month

Commercial notes

Phoenix open source is a genuine free path with no feature gates for teams that self host. The paid path runs from $50 Pro to a custom Enterprise contract, with compliance (SOC2 Type II, HIPAA), audit logs, enterprise SSO, dedicated support, and adb Data Fabric reserved for Enterprise.

Key ambiguities

Total cost at scale depends on monthly span volume and ingestion, and on whether compliance needs force a move to the custom Enterprise tier.

Cancellation / refund

AX Free and AX Pro are self serve and month to month, and Enterprise terms are contractual. No cancellation or refund terms are published.

Support SLA / resale

Community support on Free, email support on Pro, dedicated support with a contractual uptime SLA on Enterprise.

Missing data

Enterprise pricing is custom and not listed. Startup pricing is available by application.

Agentic Index verified 2026-09-21

Alternatives to Arize AI

The closest documented capability profiles to Arize AI among agent infrastructure platforms tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.

Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.