July brought two platform releases, 2026.28 and 2026.30, introducing a genuine data analysis toolkit inside chat, a more polished and governable Agentic Table, and a wider choice of AI models. Here's what's new, and what it means for your day-to-day work.
A Genuine Data Analysis Toolkit, Built Into Chat
You can now ask your assistant to actually run the numbers, not just describe them. The new Code Execution tool is generally available and lets models write and run Python code in a secure, containerized environment — so requests involving math, data analysis, or multi-step calculations get precise, code-based answers instead of estimates. Whether you're reconciling figures, modeling a scenario, or working through a data set, the assistant can now do the calculation itself and show its work.
A More Polished, Governable Agentic Table
Agentic Table keeps maturing as a serious due-diligence tool. Comments are now generally available on every sheet.
.png?width=2160&height=953&name=July%20-%20Leaving%20comments%20in%20agentic%20table%201%20(2).png)
.png?width=2160&height=953&name=July%20-%20Leaving%20comments%20in%20agentic%20table%202%20(1).png)
.png?width=2160&height=953&name=July%20-%20Leaving%20comments%20in%20agentic%20table%203%20(1).png)
The copy action is now available to everyone in the Answers columns, and error messages are clearer across the board.
.png?width=2160&height=953&name=July%20-%20Copy%20(1).png)
Sheet owners also get finer control over collaboration: the Section and Question columns can be locked down for "can answer" users so question structure stays protected, assignees can no longer remove themselves from a sheet without permission, and administrators can now configure how the Section column is formatted per space — useful as more teams share sheets across departments.
Expanded Access to the Latest AI Models
Your model options just grew. Claude Sonnet 5, GLM 5.1 and 5.2, and the GPT-5.6 family (Sol, Terra, and Luna) are all now available on the platform, giving you more flexibility to match a model to the task at hand. A handful of older models were retired on July 10, and any workspace still using Claude Opus 4.1 has already been migrated to Opus 4.8 ahead of its own retirement, no action is needed on your end.
For Platform Administrators
Administrators get more visibility into the knowledge base this month. New ingestion metrics track file counts, additions and removals, and upload channel over time, and a self-service embedding-model switch lets you change a tenant's embedding model, including dimension changes, without manual intervention. Analytics reports also now capture which tools and skills were used in a given conversation, making it easier to understand how your team is actually using the platform.
As always, these updates roll out automatically — there's nothing for your team to install. If you'd like a walkthrough of any of these features, or want to enable one that isn't yet active in your workspace, reach out to your customer success contact.
What's new in Unique AI Platform in July 2026
July brought two platform releases, 2026.28 and 2026.30, introducing a genuine data analysis toolkit inside chat, a more polished and governable Agentic Table, and a wider choice of AI models. Here's what's new, and what it means for your day-to-day work.
A Genuine Data Analysis Toolkit, Built Into Chat
You can now ask your assistant to actually run the numbers, not just describe them. The new Code Execution tool is generally available and lets models write and run Python code in a secure, containerized environment — so requests involving math, data analysis, or multi-step calculations get precise, code-based answers instead of estimates. Whether you're reconciling figures, modeling a scenario, or working through a data set, the assistant can now do the calculation itself and show its work.
A More Polished, Governable Agentic Table
Agentic Table keeps maturing as a serious due-diligence tool. Comments are now generally available on every sheet.
The copy action is now available to everyone in the Answers columns, and error messages are clearer across the board.
Sheet owners also get finer control over collaboration: the Section and Question columns can be locked down for "can answer" users so question structure stays protected, assignees can no longer remove themselves from a sheet without permission, and administrators can now configure how the Section column is formatted per space — useful as more teams share sheets across departments.
Expanded Access to the Latest AI Models
Your model options just grew. Claude Sonnet 5, GLM 5.1 and 5.2, and the GPT-5.6 family (Sol, Terra, and Luna) are all now available on the platform, giving you more flexibility to match a model to the task at hand. A handful of older models were retired on July 10, and any workspace still using Claude Opus 4.1 has already been migrated to Opus 4.8 ahead of its own retirement, no action is needed on your end.
For Platform Administrators
Administrators get more visibility into the knowledge base this month. New ingestion metrics track file counts, additions and removals, and upload channel over time, and a self-service embedding-model switch lets you change a tenant's embedding model, including dimension changes, without manual intervention. Analytics reports also now capture which tools and skills were used in a given conversation, making it easier to understand how your team is actually using the platform.
As always, these updates roll out automatically — there's nothing for your team to install. If you'd like a walkthrough of any of these features, or want to enable one that isn't yet active in your workspace, reach out to your customer success contact.