27 Observability & Evals AI tools
AI services in the Observability & Evals field. Click a name to see its description, pricing and alternatives.
Observability & Evals: free vs paid at a glance
13 tools with official pricing data. 12 have a free plan, 0 offer a Korean UI. Cheapest paid plan: Opik ($19/mo).
| Tool | Free | Cheapest monthly | Korean | Where |
|---|---|---|---|---|
| AgentOps | Yes | $40/mo | No | Web, API |
| Arize AI | Yes | $50/mo | No | Web, CLI |
| Braintrust | Yes | $249/mo | No | Web, CLI, API |
| Evidently | Yes | — | No | Web, API |
| Fiddler AI | Yes | — | No | Web, API |
| Helicone | Yes | $79/mo | No | Web, API |
| Langfuse | Yes | $29/mo | No | Web, CLI, API |
| Opik | Yes | $19/mo | No | Web, API |
| Phoenix (Arize) | Yes | $50/mo | No | CLI |
| Promptfoo | Yes | — | No | CLI, Web |
| PromptLayer | Yes | $49/mo | No | Web, API |
| Traceloop | Yes | — | No | Web, API |
| MMLU | No | — | — | — |
From each tool's official pricing page (checked by 2026-10-14). '—' means not stated. Prices change, so check before paying.
How to choose
- Try a tool with a free plan first and check the limit before paying.
- If you need Korean menus, filter by the Korean column.
- Annual plans are often cheaper; the table shows monthly prices.
AgentOps
Observability and debugging platform purpose-built for AI agents
Read the guide →Free + paidArize AI
ML observability and monitoring platform
Read the guide →Free + paidBraintrust
LLM evaluation and prompt management platform
Read the guide →Free + paidCaptum
Model interpretability library for PyTorch
Read the guide →Free + paidDeepEval
Open-source LLM evaluation framework
Read the guide →Free + paidEleutherAI LM Evaluation Harness
Framework for few-shot language model evaluation
Read the guide →Free + paidELI5
Debug and explain ML classifiers and regressors
Read the guide →Free + paidEvidently
ML model monitoring and data quality checks
Read the guide →Free + paidFiddler AI
ML model performance management
Read the guide →FreeGuidance
Constrained generation and prompt control by Microsoft
Read the guide →FreeHelicone
Open-source LLM observability and proxy
Read the guide →Free + paidHumanEval
OpenAI's benchmark for evaluating code generation
Read the guide →Free + paidInterpretML
Microsoft's toolkit for model interpretability
Read the guide →FreeLangfuse
Open-source LLM engineering platform for tracing and evals
Read the guide →Free + paidLangSmith
Platform for debugging and evaluating LLM apps
Read the guide →Free + paidLIME
Local interpretable model-agnostic explanations
Read the guide →FreeOpik
Open-source LLM evaluation and tracing by Comet
Read the guide →FreePhoenix (Arize)
Open-source AI observability and evaluation platform
Read the guide →FreePromptfoo
Test, evaluate, and red-team LLM prompts
Read the guide →Free + paidPromptLayer
Prompt versioning and management platform
Read the guide →Free + paidRagas
Evaluation framework for RAG pipelines
Read the guide →Free + paidSHAP
Game theoretic approach to explain ML model outputs
Read the guide →Free + paidTraceloop
OpenTelemetry-based LLM observability
Read the guide →Free + paidHELM
Holistic evaluation of language models
View on the 3D map →Free + paidMMLU
Massive Multitask Language Understanding benchmark
View on the 3D map →Free + paidPromptFlow
Suite of development tools for LLM apps
View on the 3D map →Free + paidWhyLabs
AI observability and data monitoring
View on the 3D map →