Side by side

Arize Phoenix vs DeepEval

Both stand on llm developer tools. Which one you can start free with, which one lets you read the code, and what each costs once you pay.

Arize Phoenix

Free plan APISelf-hostedWeb

Phoenix brings tracing, evaluation, prompt inspection, and dataset experiments to AI application teams. It is free and open source, with deployment choices and usage management left to users.

Visit Arize Phoenix

DeepEval offers a broad toolkit for evaluating AI systems across modalities, metrics, and agent workflows. Its open-source edition is positioned for pre-production testing; enterprise collaboration and deployment options are described separately through Confident AI Evals.

Visit DeepEval

The facts

	Arize Phoenix	DeepEval
Free plan	Yes	Yes
Open source	Not published	Yes
Runs on	API, Self-hosted, Web	API, Linux, Mac, Self-hosted, Web, Windows
Lowest paid plan	Free or on request	Starter: $200 a month
Plans published	1	5
Maker	Not published	Not published
Website	arize.com	deepeval.com
On llm developer tools	#5 of 26	#1 of 26

Arize Phoenix: what it does well

Traces model calls, retrieval, tool use, and custom logic.
Evaluations can use models, code checks, or human labels.
Prompt tools include versioning, comparison, and replay.
Runs locally or through several self-hosted options.
Free and open source, with user-managed usage limits.

What to know first

Users manage trace spans, ingestion, projects, and retention.
Self-hosted authentication requires a compatible OIDC endpoint.

DeepEval: what it does well

Pytest-native evaluations run in CI/CD
Lists more than 50 evaluation metrics
Evaluates text, images, and audio
Supports synthetic test data and persona simulations
Integrates with agent frameworks and model providers

What to know first

DeepEval OS is limited to pre-production testing
Default telemetry is sent to PostHog
OS results are kept in local files

Alternatives to Arize Phoenix Alternatives to DeepEval The whole list