Experimental Speedometer workloads can be used for workloads that do not directly meet the criteria for official workloads:

• Incubation of new workloads
• Edge-cases that don't work well in the official setup
• Testing features that aren't fully shipped in all browsers

The main purpose is to have a space for low-friction contributions to experiment with new ideas. There is no explicit requirement to promote experimental workloads to the official selection.

Contributions to experimental workloads can be merged with the approval from a single vendor.

This is the opposite from official workloads that are scored, where we require approval from all vendors.

• Stored in the /experimental folder
• Disabled by default and thus do not contribute to the official score
• Use the experimental tag in the Suites declaration

Experimental workloads will not be officially hosted on Speedometer releases on https://browserbench.org, unlike the official workloads.