OpenAI Evals

More in Evaluation

• DeepEval

  Open-source LLM evaluation framework. Unit-test your LLM outputs with 14+ metrics including hallucination, toxicity, and RAG.

• Promptfoo

  Test and evaluate LLM prompts and RAG pipelines. Red-teaming, A/B testing, and CI integration for prompt quality.

• AgentBench

  Benchmark for evaluating LLMs as agents across 8 distinct environments including OS, web browser, and database tasks.

• Label Studio

  Open-source data labeling platform. Annotate text, images, audio, and video for training and evaluating AI models.

• Ragas

  Framework for evaluating RAG pipelines. Measures faithfulness, answer relevancy, context recall, and more without ground truth.

• Argilla

  Open-source data curation platform for LLMs. Collect human feedback, curate training data, and evaluate model outputs collaboratively.