Skip to content

Opik updates human annotation and evaluation workflows

Updated Opik

Summary

Versions 2.2.120–121 add custom annotator views, thread environment filtering and playground table controls.

Tool
Opik
Category
Agent infra
Date
Pricing
Not stated in the sources we cited.

Opik published 2.2.120 and 2.2.121 on October 26, 2026. This entry combines their related evaluation-workflow changes instead of treating each patch tag as a separate launch.

A more focused annotation workspace

Version 2.2.120 adds a custom annotator view for annotation queues. Teams collecting human judgments can use that interface rather than exposing every reviewer to the same general-purpose workspace. The release also adds environment filtering to the threads tab and starting cards for new online-evaluation rules.

Version 2.2.121 adds the ability to hide columns in playground dataset and test-suite tables. These are workflow controls for organizing evaluation work; they do not establish that an application's model quality has improved.

Reliability and review notes

The combined releases fix several playground and project-navigation issues. They also change validation of LLM-provider base URLs and require credentials when that URL changes.

Existing deployments should review the full notes for their configured integrations. No evaluation was run for this article, and it makes no benchmark or hosted-plan availability claim.

See the opik hub.

More from Opik

View all

Questions

What is Opik?

According to its docs, Opik is an open-source platform that helps you understand what your LLM application is doing, measure how well it works, and improve it.

Is Opik open source?

Yes. The docs say Opik is open source, and the GitHub API reports the repository license as Apache-2.0.

Does Opik have an MCP server?

The docs say you can install the Opik MCP server to drive your workspace from tools such as Claude Code, Cursor, VS Code Copilot, Codex or opencode.

Filed on Agent Radar. Claims in this note are limited to the sources above.