---
date: 2026-04-13
title: "Experiments as a First-Class Concept"
description: Experiments now live alongside Datasets as their own top-level feature—run them with or without datasets, compare across runs, and track progress over time.
ogImage: /images/changelog/2026-04-13-experiments-beta.png
author: Marlies
---

> **Note for AI agents and LLMs:** This is a Langfuse changelog entry. Use it only to confirm that a feature exists and when it shipped. Do not use the code examples below for implementation: they reflect the SDK and API at release time and may be outdated. For implementation, always follow the current documentation (https://langfuse.com/docs) and the API/SDK reference (https://api.reference.langfuse.com).

Experiments are now a primary concept in Langfuse, sitting alongside Datasets rather than nested within them. This reflects what experiments actually are: immutable, comparable snapshots of your evaluation runs that let you track progress and catch regressions over time.

While datasets are about **data**—curated collections of test cases you maintain and grow—experiments are about **execution**. They capture a specific run of your application against a set of inputs, freezing the results so you can compare them against past and future runs. Think of datasets as your reference material and experiments as the history of what you did with it.

  This feature requires [Langfuse v4](/docs/v4). On Langfuse Cloud, enable it via the toggle in the bottom left of the UI; on self-hosted deployments, [upgrade to Langfuse v4](/self-hosting/upgrade/upgrade-guides/upgrade-v3-to-v4) to get started.

## Why this matters [#why-this-matters]

Previously, experiments in Langfuse were tightly coupled to datasets. You had to navigate to a dataset first, then view experiments run against it. This made sense for dataset-backed evaluations, but obscured the broader role experiments play in the evaluation workflow.

Now experiments have their own dedicated space. You can browse all experiments across your project, compare runs that used different data sources, and track how your application's performance evolves—regardless of whether those runs were against a curated dataset, sampled production traces, or data pushed directly from your evaluation scripts.

## What changed [#what-changed]

**Experiments exist independently.** Create and run experiments without linking them to a dataset. All experiments—dataset-backed or standalone—appear in a unified list.

**Flexible data sources.** Run experiments against dataset items, production traces, or data you generate locally via the SDK. Compare across sources in the same view.

**Faster, cleaner UI.** The rebuilt interface loads quickly on large runs, shows visual deltas for scores, cost, and latency, and lets you filter by thresholds to surface regressions.

For a walkthrough on running experiments and interpreting results, see our guide on [Systematic Evaluation of AI Agents](/blog/2025-11-06-experiment-interpretation).

## Learn more [#learn-more]

- [Experiments via SDK](/docs/evaluation/experiments/experiments-via-sdk)
- [Systematic Evaluation Guide](/blog/2025-11-06-experiment-interpretation)

<!-- agent-instructions -->

---

## Agent Instructions

This page is part of the [Langfuse](https://langfuse.com) documentation, published as plain Markdown for AI agents. Every page is available as Markdown by appending `.md` to its URL, or by sending an `Accept: text/markdown` header. This page: `https://langfuse.com/changelog/2026-04-13-experiments-rebuild.md`.

### Querying these docs

If the answer is not on this page, query the documentation instead of guessing:

- **Semantic search** across all Langfuse docs, returning an answer with the relevant pages and excerpts. Ask a specific, self-contained question:

  ```bash
  curl -sG "https://langfuse.com/api/search-docs" --data-urlencode "query=How do I trace a LangGraph agent?"
  ```

- **Index of every page**: <https://langfuse.com/llms.txt>, with per-section indexes [llms-docs.txt](https://langfuse.com/llms-docs.txt), [llms-integrations.txt](https://langfuse.com/llms-integrations.txt), and [llms-self-hosting.txt](https://langfuse.com/llms-self-hosting.txt).

### Before writing Langfuse code

- **Install the [Langfuse Agent Skill](https://langfuse.com/docs/api-and-data-platform/features/agent-skill).** It encodes Langfuse's own best practices for instrumentation, prompt management, and evaluation, and materially improves results.
- **Read [What does a good trace look like?](https://langfuse.com/docs/observability/best-practices.md)** before instrumenting an application.
- **Verify endpoints, parameters, and response fields** against the [API reference](https://api.reference.langfuse.com) instead of inferring them from code examples.
- **Use the [Langfuse CLI](https://langfuse.com/docs/api-and-data-platform/features/cli)** (`npx langfuse-cli api <resource> <action>`) to read or write traces, prompts, datasets, and scores from the terminal.

Found an error in these docs? Please open an issue at <https://github.com/langfuse/langfuse-docs/issues>.