Welcome to Volume 5 of our 2026 product release notes! We've been working hard to bring you performance improvements that add value to your work. Read on to learn about the product enhancements and bug fixes released in May 2026.
Some features in this release come from Benchling's annual Polish Bash ⭐ the event where our product team focuses on smaller updates that refine and sharpen Benchling.
Benchling AI
Benchling's artificial intelligence agents and features.
New Benchling AI chat experience[2026.5.AI.1]
This update brings together all the capabilities of our previous agents into a single, more powerful agent. With one unified interface you can import data, create charts and presentations, write reports, analyze data, and more
As part of this update, the AI settings in the Tenant Admin Console have been reorganized for better clarity.
The Benchling MCP server supports the new Benchling AI chat experience [2026.5.AI.2]
You can query the new Benchling AI chat experience, directly from the Benchling MPC server, by using the “Benchling agent” call. Admins can enable the Benchling MCP for their tenant via a new setting in the AI Settings under the Tenant Admin Console.
New Agent skills capabilities now supported in Benchling [2026.5.AI.3]
Agent Skills are now available for Benchling AI agents. A variety of Benchling-authored default skills are included by default. Users can add their own custom skills by uploading a skill .MD file, or by using Benchling AI to create the skill for them.
- Using the gear icon within the Benchling AI chat interface icon, available skills can be viewed by each user. The skill’s contents can be expanded and viewed, including the skill .MD file. From this menu, skills can be toggled on and off depending on user needs using the. By default, a skill is toggled on.
- User created skills can be both created and edited using Benchling AI. When a new version of the skill is created, the agent preserves the prior version history for audit and rollback.
- Skills can be utilized by default with the new Benchling AI chat experience (launched June 3, 2026). When using legacy agents, skills are compatible with the Deep Research and Compose Agents.
Benchling AI Skill: Data Analysis [2026.5.AI.4]
Move from data to insight in minutes. Benchling AI analyzes your experimental results — running statistical tests, fitting models, generating visualizations, and surfacing batch effects — so you can interpret findings without leaving your notebook. Accessible directly from the chat interface.
Benchling AI Connectors is now GA [2026.5.AI.5]
AI Connectors can now be configured to give Benchling AI access to data that is stored in external systems. A directory of AI Connectors is supported out-of-the-box and custom AI Connectors can be set up for any other system that supports the Model Context Protocol. See all currently available connectors.
- Admins can install new AI Connectors for the tenant in the AI Settings tab of the Tenant admin console. After an admin installs a connector, users can access it in their personal AI Chat Settings.
- Previously some capabilities available from AI Connectors were referred to as MCPs within the product.
Improvements to models within Model Hub [2026.5.AI.6]
- AlphaFold2 now runs on faster GPUs, reducing time from request to structure prediction result.
- Added support for multimers on OpenFold3, available via the 3D structure tab on AA sequences.
The Benchling UI has been updated for clarity [2026.5.AI.7]
- Buttons to access Benchling AI from the template collections, and from global create have been updated.
- The spacing of text within Custom instruction preview text has been improved to avoid text being overly truncated.
- The steps shown within the chat interference while the agent is processing, now displays thinking steps differently when a built-in skill vs a user-created skill is used.
[BETA] Model Hub API available in V3 Alpha API [2026.5.AI.8 ]
Users/tenants that subscribe to the V3 and Alpha api endpoints can access Model Hub's new API endpoints for creating and viewing prediction jobs.
Notebook
Benchling Notebook helps users log their experiments, create protocols, and capture results against their samples.
Multiple lookup columns can be created at once [2026.5.N.1]
Users can now create multiple lookup columns in a single action by selecting multiple lookup fields.
Freeze table columns [2026.5.N.2]
Users can now set how many columns to "freeze" on a table's left-hand side to remain visible while scrolling horizontally.
Display all cell statuses on hover [2026.5.N.3]
Structured Table cells will now display all statuses (Error, Unsubmitted, Outdated, etc.) on hover instead of only the highest priority status.
Print notebook entries with formulas displayed [2026.5.N.4]
Users can now print notebook entries with table formulas visible when the Show Formulas toggle is enabled. Previously, printed views only displayed calculated values regardless of the toggle state.
- Polish Bash item! ⭐
Sticky horizontal scrollbar for wide tables [2026.5.N.5]
The horizontal scrollbar for wide tables will now follow the user's window. Users can now use the horizontal scrollbar without first scrolling to the bottom of the table.
- Polish Bash item! ⭐
Minor versioning now possible with template publishing [2026.5.N.6]
Template authors can now optionally classify each published version as minor. When publishing, there's now a minor version checkbox. If it is not selected, the system by default auto-numbers to the next whole number (e.g., Version 3). If it is selected, it auto-numbers to a sub-version (e.g., Version 2.1, 2.2). Version type shows up in version history so reviewers and auditors can quickly distinguish substantive changes from typo fixes
- Polish Bash item! ⭐
Molecular Biology
Benchling's Molecular Biology is a cutting-edge in silico design tool that supports the latest scientific workflows.
Benchling Biologics [2026.5.MB.1]
Benchling Biologics is an end-to-end platform that gives antibody R&D teams the structured, AI-ready data foundation they need to move faster: an antibody-aware data model, no-code format configuration, and automated registration that connects sequences to proteins to experimental context without the data wrangling.
- This is a new offering that requires a commercial agreement to enable.
- Learn more
Users can create multiple translation regions at once [2026.5.MB.2]
Users can now create multiple translation regions in a single action from the sequence editor.
In Vivo
Manage in vivo studies with modern software designed for data capture, collaboration, and insights.
Display folder name in settings and study overview [2026.5.IV.1]
The selected folder location for the In Vivo Sync now displays the folder name after selection, rather than only the folder ID, when registering this study and its associated entities.
Saved table views [2026.5.IV.2]
Each user can now save up to 10 views per table, capturing their preferred column visibility, filters, sort order, column widths, and rows per page. Saved views persist across sessions, and a view configured in one study automatically applies to that same table type everywhere that user accesses it.
- For access to this feature, contact: Benchling Support
Sync error events resolve after successful retry [2026.5.IV.3]
Sync error events now back-up date to resolve when a later sync succeeds for the same item.
Refreshed navigation bar in Benchling In Vivo [2026.5.IV.4]
The navigation bar in In Vivo studies was updated to consolidate the User Account, Manage team, Help, and Log out options in a user's menu. The quick add button was also removed.
Cancelled status available for tasks [2026.5.IV.5]
Tasks now support a Cancelled status, so that when they are incomplete, they can be closed out instead of going overdue.
Workflow displays 0% body weight change [2026.5.IV.6]
Workflow now displays a 0% body weight change value when applicable instead of hiding it.
Condition for task generation displayed on task [2026.5.IV.7]
Users can now see which condition triggered a task's generation on a given date. When multiple conditions trigger the same task, the assignment reason is shown per animal in the task's target section.
In Vivo supports unknown sex for animals. When syncing animal Sex into a dropdown field on the core tenant Animal Schema, an option called "unknown" must exist in this dropdown to avoid sync errors. [2026.5.IV.8]
When this setting is enabled, Sex includes a third option, 'Unknown', for cases where an animal's sex hasn't yet been determined.
- For access to this feature, contact: Benchling Support
Allow study owner access to study draft [2026.5.IV.9]
When a study author registers a study on behalf of someone else and identifies that person as the Study Owner, the named owner now has access to the draft. This means the author and the owner can collaborate on the study design together before the study is finalised.
- Polish Bash item! ⭐
Registry and Inventory
The Registry and Inventory combine to form an extremely powerful solution for life sciences R&D samples. The Registry can be used to model and log any biological entities while the Inventory provides an interactive digital window into the physical storage system of your lab.
Protein registration now strips out whitespace in DNA and AA inputs instead of erroring [2026.5.RI.1]
DNA sequence and AA residue inputs now ignore whitespace characters when validating and creating sequences, preventing errors from pasted strings.
Parent link fields can now point to a fieldset [2026.5.RI.2]
Parent links to fieldsets are now generally available for all users.
[BETA] Added 3 new endpoints in v3 container endpoints [2026.5.RI.3]
Added 3 new beta endpoints in v3 container endpoints to provide parity with V2 APIs. Existing container update endpoints are also updated to allow check in/check out status updates.
Workflows
Workflows are a sequence of processes that are performed to answer a research question or synthesize new material.
Default values for tasks created from structured tables [2026.5.W.1]
Default values are selected from the left most column of a structured table when creating tasks from a structured table.
Ability to add teams/groups as workflow watchers [2026.5.W.2]
Users can now add teams or groups as watchers to workflow tasks
Insights
Insights enables users to query, visualize, and share high quality, structured data that resides in Benchling.
Automation Designer
Automation Designer is Benchling's no-code/low-code workflow builder that extends the Connect and Analysis products, enabling users to design and orchestrate multi-step workflows, from instrument runs and data transformations to downstream analysis, within a single, connected environment.
Access to Automation Designer: Access is included for customers with Advanced Analysis and Connect enabled on their account. For more information speak to your Benchling account team.
Data Entry Agent (DEA) step in Automation Designer [2026.5.IN.1]
The Data Entry Agent (DEA) step in Automation Designer brings AI-driven data extraction directly into structured automation workflows. Users can now embed a DEA step alongside pipeline steps such as Files to extract, map, and transform variable or unstructured inputs, such as CRO-generated PDFs or raw instrument outputs, before passing results into downstream deterministic steps like chart creation, results saving, or notebook export. This hybrid model combines the flexibility of AI with the consistency and reusability of a structured data pipeline, keeping the full workflow within Benchling and preserving data traceability end to end.
Key capabilities:
- Embed a Data Entry Agent step directly within an Automation Designer, including alongside the Files step for document-based extraction workflows
- Use AI to handle non-deterministic extraction and transformation tasks, mapping fields, parsing variable formats, and pulling structured outputs from unstructured inputs
- Pass DEA outputs into downstream deterministic data pipeline steps for consistent, repeatable processing run-over-run
- Preview AI-generated dataset results before executing the full pipeline step, displaying up to 5 rows for quick validation
Box Plot GA [2026.5.IN.2]
Box Plot adds a box-and-whisker chart type to Insights Analysis, enabling scientists to visualize distributions and make group-level comparisons directly within the platform
Format numeric columns with decimal precision [2026.5.IN.3]
Users can now format numeric columns in analysis pipelines by specifying decimal precision from 0 to 20 places. This transformation applies consistent number formatting across datasets for improved readability and standardization.
Analysis Left-Sidebar Improvements [2026.5.IN.4]
Several improvements to the left sidebar in the Analysis module have been added to make it easier to organize, find, and navigate views.
- Views can now be nested more than one level deep, giving users greater flexibility to organize complex analyses in a hierarchy that matches a user's workflow.
- A new search experience lets users quickly find any view within the sidebar without having to scroll through long lists, to the right view faster.
- Users can also now create new views directly from any parent or child view in the sidebar, reducing the steps needed to build out an analysis.
- Clicking a view in the sidebar now focuses it in the Flowchart, making it easy to navigate between a sidebar and the visual flowchart.
- Users can now collapse or expand all sections in the Analysis sidebar with a single action.
v3-alpha API supports editing Insights dashboard queries [2026.5.IN.5]
The v3-alpha API now supports editing Insights dashboard queries.
Benchling Connect
Benchling Connect, formerly Lab Automation, enables seamless integration of Benchling to liquid handlers, plate readers, imaging instruments, and more.
Optional ASM vendor auto-detection [2026.5.BC.1]
The Convert Instrument Files to ASM step of the Automation Designer now supports auto-detection without requiring an ASM vendor value.
Connection Schemas no longer have access policy configurations [2026.5.BC.2]
The access policies tab has been removed from Connection Schemas, as it doesn't provide additional functionality to users when permissions were granted
Update Allotropy to 0.1.127 [2026.5.BC.3]
Benchling's integration with Allotropy has been upgraded to version 0.1.127
- See the: Change Log
Schemas
Schemas are the structural foundation for how your information is categorized in Benchling. They are categories created specifically for your tenant that can represent materials, data, storage methods, and tasks.
Automatic creation of Fieldset fields supported in schemas create API [2026.5.SCH.1]
The schemas CREATE API now supports applying a fieldset and ensuring the fieldset fields and properties are automatically configured on the schema.
An API resource for Fieldsets has been added to the V3 Alpha API [2026.5.SCH.2]
Fieldset configuration data can now be created, updated, and read through the V3 Alpha API. Users can interact with Fieldsets in the same way they interact with them in the UI. Customers can create/update Fieldsets, set prefixes, and configure fields.
Entity Schemas can be duplicated [2026.5.SCH.3]
Users can now duplicate an existing Entity Schema to create a new one without rebuilding it from scratch.
The Schemas and Fieldsets tab hides archived assets [2026.5.SCH.4]
The Schemas and Fieldsets tab of a dropdown now only displays schemas and fieldsets where the dropdown field is active, hiding entries from archived fields.
Archive button added to schema settings pages [2026.5.SCH.5]
Archive and unarchive buttons are now available directly on individual assay schema settings pages, including Run Schemas, Result Schemas, Legacy Request Schemas, and Task Schemas, eliminating the need to navigate to listing pages.
Inherited System Type fields visually distinguished on schemas [2026.5.SCH.6]
Schema configuration pages now visually distinguish fields inherited from System Types (e.g. Serial number on Equipment types) from fields defined directly on the schema and fields added by a Fieldset.
Schema settings shows consistent duplicate button [2026.5.SCH.7]
Update Task, Entity, and Run Schema setting pages to render a consistent "Duplicate" button next to the "Export" dropdown that includes audit log exports.
[BETA] A schema resource for each Entity Schema type is now available in the V3 Alpha API [2026.5.SCH.8]
The following API resources have been added to the V3 Alpha API: CustomEntitySchema, AaSequenceSchema, DnaOligoSchema, DnaSequenceSchema, OligoConjugateSchema, OligoDuplexSchema, RnaOligoSchema, RnaSequenceSchema, MixtureSchema, MoleculeSchema, EquipmentSchema
These new API resources support all of the same behaviors of Entity Schemas in the UI with the exception of naming templates and constraints.
Users can create/update schemas, set prefixes, configure fields and apply fieldsets.
[BETA] Beta stability API resources have been added [2026.5.SCH.9]
Beta stability API resources for Unit and UnitType writes have been added to the V3 API.
[BETA ]Replace DropdownFieldDefinition with DropdownLinkFieldDefinition in V3 API [2026.5.SCH.10
Removes DropdownFieldDefinition. It has been replaced by DropdownLinkFieldDefinition, which is identical except it replaces the deprecated "dropdown" field with "linkDefinition" to be consistent with other field definition objects.
Access Management
Access management is the way customers control who is able to access which resources, and includes things like project permissions, access policies, and admin roles.
Delegated auth is now GA and available on all tenants [2026.5.AM.1]
Delegated Auth allows apps to act on behalf of individual users, so the app is limited to what that user can see/do, and any actions the user takes via the app are attributed to the user in the audit log or other fields like created by. The audit log still has information about which app was used to make those changes so end-users will have a complete picture of who made which changes via what mechanism.
Application Core
To create a seamless Benchling experience, our platform must be consistent, compliant and predictable. The Application Core provides the central capabilities to create a smoother user experience.
Ordering from Twist, Gingko, and Adaptyv in Benchling [2026.5.AC.1]
Benchling now supports direct ordering from our partners Twist, Adaptyv, and Ginkgo. Scientists can design candidates, place orders for synthesis and services, and receive structured results back without ever leaving Benchling.
- Learn more. To gain access to this capability, contact your Benchling Representative.
Clearer audit export column names [2026.5.AC.2]
Columns in the export and in-app audit log for user name and user handle now use clearer labels
Audit log timestamp column now sortable [2026.5.AC.3]
The timestamp column in the in-app audit log modal now supports interactive sorting. Users can toggle between ascending and descending order by clicking the column header, making it easier to find recent or historical audit events.
Audit log displays formatted JSON values [2026.5.AC.4]
The in-app audit log now automatically detects JSON in Old Value and New Value columns and displays it as formatted, indented text for easier reading.
New configuration migration UI is now GA [2026.5.AC.5]
The new configuration migration UI which uses tabs instead of steps is now GA for all tenants. This change allows customers to choose which order they'd like to manage their objects in config migration, and also makes the process more efficient because validation only has to be done at the end instead of being re-run at every step. The old UI will remain temporarily available, but will be removed in a future release
Attachments with Exported Entities[2026.5.AC.6]
Added support for optionally including file attachments when exporting entities (DNA sequences, molecules) as ZIP. When the file limit is reached, remaining attachments are skipped
An ATTACHMENTS_NOTICE.txt file is added to the ZIP explaining which attachments were omitted and why.
- For access to this feature, contact: Benchling Support
- Polish Bash item! ⭐
Changed default view of Team in the Tenant Admin Console [2026.5.AC.7]
Now when clicking into the team view within the tenant admin console, the default view will be the "Members" tab.
- Polish Bash item! ⭐
Compliance Controls
With Compliance Controls, gathering required information for compliance is simple, intuitive, and performant. The team's primary focus is on our auditing platform but we also own exporting and archiving.
Project name in export failure emails [2026.5.CC.1]
Export failure emails now include the project name to help users identify the affected resource.
Failing exports retry automatically with a different layout [2026.5.CC.2]
When a large entry is exported, Benchling may use an alternate layout to ensure the export succeeds.
- For access to this feature, contact: Benchling Support
Exports now show an In Progress status when a review has not yet been initiated. [2026.5.CC.3]
Exports now show an In Progress status when a review has not yet been initiated.
Export layout notice for large entries [2026.5.CC.4]
When a large entry is exported, Benchling may use an alternate layout to ensure the export succeeds. In these cases, the export's metadata now includes a note letting a user know an alternate layout was used.
Developer Platform
Our developer platform is fundamental to centralizing and standardizing all your R&D data. Benchling's APIs and Events are built to match the flexibility and speed of modern life science R&D.
Home panel now includes status of long-running tasks [2026.5.DP.1]
The status of longer running operations like chat queries, structure predictions, and protein creations is now included in the home panel.
[BETA] Beta stability v3 APIs for audit logs now available [2026.5.DP.2]
The v3 Audit Log API includes four new endpoints at /api/v3-beta/benchling/audit-log/ that return audit entries as structured JSON with stable per-entry IDs (auditeff_ prefix), rich filtering (date range, user, action, reason code), sorting, pagination, selective field return, and multi-object queries (up to 10 objects per request). CSV/PDF export is also available through a dedicated v3 async custom method.
[BETA] More bulk endpoints in v3-beta [2026.5.DP.3]
Previously, all bulk write actions in v3 were limited to v3-alpha. With this update, bulk endpoints can be promoted to beta or stable when appropriate. Most notably, this update includes the promotion of the Results bulk create and update endpoints to v3-beta.
[BETA] Added listing (GET) endpoints for WellPlatePosition in V3 API [2026.5.DP.4]
Users can now list wells and query wells by their api id from v3 APIs.
[BETA] New batch endpoints in v3 REST API [2026.5.DP.5]
The v3 REST API now includes batch APIs for all write actions. With these APIs, up to 25 objects can be created or modified in one operation. These endpoints are an alternative to our bulk APIs that handle larger scale write operations.
BioProcess
Our BioProcess offering helps with process development studies, including non-In Vivo workflows. It includes tools for tracking, managing, and analyzing experiments like process development studies, recipes, and studies.
Improved template collection navigation from recipes [2026.5.BP.1]
Template collection links on recipes now navigate to the template collection in the workspace navigation instead of feature settings.
- Polish Bash item! ⭐
PipeBio
Benchling's cloud-based bioinformatics software for antibody/peptide screening & drug development. Conduct BCR, TCR, VHH, scFv, or peptide sequence analysis and integrate your workflows with Benchling's core platform.
Added feature to edit sequences [2026.5.PB.1]
Inline sequence editing in the table sequence viewer is now available in AWS-hosted PipeBio environments, matching the capabilities previously available only on GCP.
Added feature to overlay multiple structures [2026.5.PB.2]
Added the ability to view multiple protein structures overlaid in a single view and download the combined image.
Added feature to predict ProteinMPNN variant structures [2026.5.PB.3]
Added an option to predict and compare 3D structures for the variant sequences produced by ProteinMPNN.
Bug Fixes
Benchling AI
- Fixed an issue where Deep Research chat sessions could become stuck in an endless thinking state, preventing users from receiving responses or starting new chats.
- Fixed an issue where agent chart generation could fail unreliably.
- Fixed an issue where connecting to MCP servers with OAuth would fail silently when dynamic client registration was blocked or unsupported
- Fixed an issue where agent chart generation could fail unreliably.
- AI Connector name can be reused after uninstall
- Fixed an issue where agent errors were not properly surfaced to the MCP server.
- Fixed an issue where the Chat agent could conflate user-authored messages with system-injected context, improving response accuracy.
Notebook
- Save when renaming tables in Recipe steps
- Formulas targeting ranges of merged cells correctly target the parent cell in merged range
- Improved Table performance with Complex Polymers
- Consistent formula resolution to chips after Table Submission
- Tab navigation stays within highlighted cells in Notebook tables
- Improved accuracy of formula complexity limits
- Entity chip fill-down copies values correctly
- Improved Results Table submission performance
- Improved archived container search in Tables
- Faster malware scanning when submitting notebook entries for review
- Multi-part file uploads complete reliably
- Fixed right-click on column and row headers
- Consistent entity chip name display in Structured Tables
- Structured Tables better prevent row limit errors
- Improved sticky table header alignment
- Fixed an issue where the formula bar would show the value from a different step when cross-step references are enabled.
- Resolved string lookup errors in Notebook
- Rolled back entry revert audit event
Molecular Biology
- GenBank export preserves structured comment formatting
- Translation regions numbered correctly on reverse strand
- NDC pipeline handles invalid germline IDs gracefully
- Rate limiting for nucleotide alignment creation
- Fusion domain errors now reported as validation errors
- Updated audit lots to use protein instead of antibody
- Assemblies API now supports skipping bins
- Bulk protein creation API now preserves distinct protein inputs
- Archived entities no longer block unlinking from DNA constructs
- Schemas with overlapping parts validation can now be unarchived
- Concurrent assembly slot updates handled correctly
- Improved missing domain errors
- Protein clone endpoint enforces read permission on source
- Stricter type validation on custom entity API
- Annotation selections clear correctly in expanded view
- Improved degenerate codon handling in sequence operations
Small Molecule
- Fixed a bug preventing users from registering multiple oligo conjugates or duplexes in a single notebook entry
In Vivo
- Fixed an issue where In Vivo Table columns did not align with the column headers on windows computers
- Planned animals validation re-runs after row removal
Registry and Inventory
- Location creation now applies application access limitations in all entry points
- Updated schema copy permissions
- Fixed a bug impacting latency when printing from Benchling via Bartender
- Bulk container creation no longer times out
Workflows
- Snapshot field transitions work in new flowchart editor
- Worklist API accepts collaborators on creation
Insights
- Analysis table replacement supports longer operations
- Fix an issue where an error was raised when creating a look up step from a top-level chart view
- Regression outputs grouped in sidebar
- Analysis steps execute reliably after refresh
- Improve handling of multi-output/single-input steps within Analysis
- Sidebar no longer errors when adding results without a join
- Adding Notebook table to Analysis table now supports more column types
Benchling Connect
- Fixed an issue for runs with an ASM conversion for Chromeleon output processing failed with a JSON parsing error when a custom transform re-triggered processing.
Schemas
- Error thrown when editing a protein schema with a fieldset link field
- Assay template draft changes no longer affect published version
- Assay template edits no longer trigger server error
- Numeric field display precision validated in the API
- Better error message when archiving a schema field through the API
- Mixture recipe unit updates now resolve legacy unit symbols case-insensitively
- Active fieldsets now appear on the Fieldsets schema page
Application Core
- Fixed an issue where the the in-app audit log would give a 500 error
Compliance Controls
- Fix entry export cutoff for Run Results tables
- Excel attachments preview correctly in validated environments
- Notebook exports no longer hang on failed lookup resolutions
Developer Platform
- Fixed issue with CMYK avatars
- POST/PATCH endpoints in the V3 API do not include creator_id as a request parameter
BioProcess
- Fix at-mentioning studies in notebook for users that do not have Bioprocess
PipeBio
- Fixed bug with job activity messages in jobs sidebar
- Fixed bug with quote handling in filter queries
- Fixed bug with handling of negative numbers in CSV export
- Fixed bug with UDF loading in jobs
- Fixed bug in import sequences dialog
- Fixed bug failing uploads on long-running jobs
- Fixed bug offering unsupported export format
- Fixed bug annotating unusual amino acids
- Fixed bug crashing sequence extraction
- Fixed bug uploading files with special characters
- Fixed bug retrieving annotations
- Fixed bug disabling wrong developability datasets
- Fixed bug showing generic differential enrichment errors
- Fixed bug breaking scroll in the jobs table