Welcome to Volume 3 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 March 2026.
Benchling AI
Benchling's artificial intelligence agents and features.
Deep Research citations improved reliability [2026.3.AI.1]
Deep Research will now only attempt to generate citations once instead of retrying on failure. This prevents unnecessary delays and ensures users receive their research reports more quickly, even if citations cannot be added.
Ask now includes inventory context for active workspace items [2026.3.AI.2]
When asking questions in Ask, the AI agent now includes inventory information, such as containers and plates, for the active workspace item. This is only available for users that have Ask access.
Create with Compose in template collections [2026.3.AI.3]
When adding a new template to a template collection, users can now select "Entry template from Compose" to draft the template content using the AI agent Compose. This is only available for users that have Compose access.
Scientific AI permissions management in admin console [2026.3.AI.4]
Tenant admins can now control access to Scientific AI models through the admin console, AI Settings tab. Permissions can be set to disabled, enabled for specific users, or enabled for all users. This gives tenant admins direct control over Scientific AI Models (previously required Benchling support to enable/disable).
Deep Research agent now performs advanced calculations with code [2026.3.AI.5]
Deep Research agents can now execute Python code to perform mathematical calculations, statistical analysis, and data transformations. This enables accurate numerical computations using libraries like numpy, pandas, and scipy for quantitative research tasks.
Agents now support switching between agent type within one chat [2026.3.AI.6]
Users can now seamlessly switch between different agents in one chat. For example start with Ask Agent, then move to Deep Research all while maintaining the same chat thread.
[Beta] MCP modal buttons follow standard placement [2026.3.AI.7]
Buttons in MCP server configuration modals now follow standard Benchling button placement conventions, with Cancel on the left and action buttons (Add, Save, Continue) on the right. This reduces the risk of accidental clicks and improves platform standardization..
[Beta] MCP tools display friendly names in chat [2026.3.AI.8]
MCP tools in chat now display user-friendly names combining the server name and tool title (e.g., "GitHub: Fetch Repository Data") instead of technical identifiers. This makes it easier to understand which tools are being used during conversations.
[Beta] Per-chat MCP server selection for Deep Research [2026.3.AI.9]
Deep Research agents can now use different MCP (Model Context Protocol) servers for each chat session. Users can select which external tool servers to enable when starting a research task, allowing customized tool access per conversation.
[Beta] MCP OAuth connection shows single success notification [2026.3.AI.10]
Previously, when connecting a new MCP server via OAuth would display multiple duplicate success notifications. The connection flow now shows exactly one green success toast when authentication completes successfully.
[Beta] MCP connection status preserved when adding servers [2026.3.AI.11]
Improved behavior so adding a new MCP server would cause all existing MCP connection status indicators to briefly reset to a checking state. Connection chips now remain stable and display the correct status while the server list updates.
Notebook
Benchling Notebook helps users log their experiments, create protocols, and capture results against their samples.
Clearer info banners for disabled editing on entries and steps [2026.3.N.1]
Info banners shown when editing is disabled on notebook entries and steps now clearly explain the reason, such as viewing a previous version or the entry being in review.
Added deleted reference column error message and check for lookups [2026.3.N.2]
Added a user facing error message when attempting to add a table with a lookup column that references a deleted column. Also added a check to prevent deleting a column referenced by a lookup in the first place
Spell check support for Notebook entry titles [2026.3.N.3]
The notebook spell check preference now applies to entry and template titles. When spell check is enabled, misspellings in title fields are highlighted automatically.
Audit trail for pasted notebook tables [2026.3.N.4]
An "Added table" audit event is now logged when a previously created structured table is pasted into a notebook entry or document. This is distinct from the existing creation audit, which is logged only when a new table is first created.
Audit trail for removed Notebook tables [2026.3.N.5]
A "Removed table" audit event is now logged when a structured table is deleted from a notebook entry or document.
[Beta] View multiple images side-by-side in Notebook entries and other documents [2026.3.N.6]
Users can now view multiple images side-by-side in horizontal alignment and grid layouts. This should only be enabled in non-production tenants while in beta.
Contact Benchling Support to enable
[Beta] Add Entry filter to V3 Results API [2026.3.N.7]
Filter results by notebook entry in V3 API
Contact Benching Support to enable
Small Molecule
The molecule entity in Registry helps users model and interact with chemical compounds throughout the platform, enabling users to draw, visualize, store, and search chemical structures.
Tenant admins can edit unit type formulas [2026.3.SM.1]
Tenant admins can now edit the formula for existing unit types in the unit dictionary UI. The base unit name and symbol remain unchanged.
In Vivo
Manage in vivo studies with modern software designed for data capture, collaboration, and insights.
New Sync error email notifications are now sent to administrators each day [2026.3.IV.1]
Administrators now receive an email each day when new Sync errors have been triggered for the In Vivo Sync integration.
Team glossaries API filter by active or archived status [2026.3.IV.2]
Team glossary list endpoints now support filtering by status (active or archived) via a new query parameter. The API response includes a status field indicating whether each glossary is active or archived, enabling display and management of glossaries based on their archive state.
Undo animal termination via API [2026.3.IV.3]
Deceased study animals can now be restored to active (living) status through the external API. Animals are reassigned to a specified cage and become available for continued study activities.
Sample plan updates now cascade to planned samples [2026.3.IV.4]
When updating sample definitions in a study, metadata changes now automatically cascade to existing planned samples. A confirmation dialog warns users about the number of affected samples before applying updates. This ensures sample metadata and details (weight/volume) remain consistent for existing samples in In Vivo and BioResearch. This does not involve creating, deleting, or cancelling samples.
Customizable Treatment labels to differentiate between similar treatments [2026.3.IV.5]
Treatments now have display labels that appear across the treatments table, workflow setup, workflow dosing, task setup, and task execution tab. Labels are auto-generated from the treatment name and dose (or volume if no dose is set), but can be edited to help differentiate similar treatments. Note that updating a treatment's name or dosing information will reset the label to the auto-generated default.
To enable contact Benchling Support.
Filter samples table by timepoint [2026.3.IV.6]
Users can now filter the samples table by timepoint using the filter dropdown. Select "Timepoint" as the filter field and choose from the available conditions: equals, is greater than, is less than, or between. The "between" option lets you specify a range (e.g. "1h" to "24h").
Mark animals as deceased Via API [2026.3.IV.7]
Animals can now be marked as deceased programatically via a new "Animal terminate" API endpoint to support integrations with colony management systems and other external tools.
[Beta] Observation task specs supported via API [2026.3.IV.8]
Task specifications with observation type can now be created, updated, retrieved, and deleted through the external API. Observation task specs support configuring observation glossary items and managing assignees and targets.
Contact Benchling Support to enable
[Beta] Measurement task specs available via API [2026.3.IV.9]
Measurement task specs can now be created, updated, and managed through the external API. This includes configuring measurement execution workflow setup, weight percentage changes, assign identifiers, and cursor positions.
Contact Benchling Support to enable
[Beta] Dosing task specs supported via external API [2026.3.IV.10]
Task specs of dosing type can now be created, updated, listed, and deleted via the external API. This enables external systems to manage dosing tasks programmatically, including treatment assignments, record new body weight setting, and skip animal workflow setup.
Contact Benchling Support to enable
[Beta] Sample task specs supported via external API [2026.3.IV.11]
Task specs with sample execution can now be created, updated, listed, and deleted via the external API. This enables external systems to manage sample tasks programmatically.
Contact Benchling Support to enable
Registry and Inventory
The Registry and Inventory combine to form an extremely powerful solution for life sciences R&D samples.
[Beta] Parent link fields can now point to a fieldset [2026.3.RI.1]
Parent link fields pointing to a fieldset enables users to specify a group of schemas that could be the parent.
Contact Benchling support to enable.
Insights
Insights enables users to query, visualize, and share high quality, structured data that resides in Benchling.
Replace columns in-place with columns configured with Add Computed Column step in Analysis [2026.3.IN.1]
Users can now replace existing columns in-place when adding computed columns, instead of creating a new column and manually hiding the original. The new column maintains the same position and display name, and downstream steps automatically reference the replaced column.
Benchling Connect
Benchling Connect, formerly Lab Automation, enables seamless integration of Benchling to liquid handlers, plate readers, imaging instruments, and more.
Integer columns now accept decimal strings (e.g "1.0") [2026.3.BC.1]
Improved integer column handling by allowing string values with zero decimal component (e.g. "1.0" -> "1") to be successfully parsed as an integer for integer dataset columns
Wider file selection modal in data retrieval [2026.3.BC.2]
File selection in Connect Run's data retrieval modal now displays at a wider size to better accommodate the file path column, improving readability when browsing and selecting files
Configurable file encoding for run input files [2026.3.BC.3]
Run input files can now be generated with UTF-16 LE encoding for instruments that require it, in addition to the default UTF-8 encoding. User can configure the encoding setting per input file specification in the run schema settings
This feature is not on by default, it is enabled by admins.
Chromeleon sequence destination override per run [2026.3.BC.4]
Chromeleon instrument connections can now override the sequence destination on a per-run basis using run fields. This allows different runs using the same connection to write results to different destinations without modifying the connection configuration each time
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.
Fieldset Entity Link Fields now clickable [2026.3.SCH.1]
Added new capability for fieldset fields with link types to schemas or dropdowns to display clickable links to those data model assets. Link fields in fieldsets now display navigation links to their referenced schemas and dropdowns.
Updated V2 Custom entity API to reflect Entity Schema types [2026.3.SCH.2]
Previously, Entity Schema types like Oligo Duplex and Oligo Conjugate were configured by creating a Custom Entity and then applying type-specific (e.g. Oligo Duplex) fields as an checkbox option. Benchling updated its user experience to show these at top-level types, meaning the user could choose to create an Entity Schema called Oligo Duplex; no longer would a user need to start with a custom entity. With that change, though, Benchling did not update our V2 API to reflect the new behavior. This update modifies the V2 custom entity write endpoints to ensure you can no longer create types like Oligo Duplex.
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.
In-App Audit Trails with Filtering [2026.3.AC.1]
View audit records directly in Benchling and filter by action, user, date range, and reason code.
Simple validation rules enabled for equipment schemas [2026.3.AC.2]
Equipment schemas now support simple validation rules like min/max values or required/disallowed substrings. Users can configure validation rules for equipment fields, matching the functionality already available for custom entities.
Warehouse database upgraded to PostgreSQL 16 [2026.2.WH.1] [2026.3.AC.3]
The Benchling Warehouse has been upgraded from PostgreSQL 14 to PostgreSQL 16. This infrastructure upgrade improves database performance and security. No changes to warehouse queries, schemas, or data access are required.
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.
[Opt-in] Custom Name for Entries in Export [2026.3.CC.1]
A custom filename template can be applied for Entry PDF exports. Custom Name filename templates are managed by Benchling Support, please reach out to have one applied to your tenant.
The following fields can be configured as part of the template:
- {ENTRY_NAME} Entry display name/title
- {CREATION_DATE} Record creation date as YYYY-MM-DD
- {ENTRY_ID} API identifier (e.g. etr_xxxxxxxxxxxx)
- {EXP_ID} Human-readable experiment ID (e.g. EXP-123)
- {DATE_SLICE} Date range of entry content (e.g. Jan 2024)
- {REVIEW_STATUS} Review status action text (e.g. Accepted)
- {REVIEW_TIMESTAMP} Timestamp of latest review status change
[Beta] Advanced Validation Rules [2026.3.CC.2]
Building on top of the foundation provided by Simple Validation Rules (released in Oct), this allows a huge amount of additional flexibility for admins to set validation that matches their scenarios. Advanced rules support using multiple fields in the same rule, referencing fields on linked entities, having conditional logic, and more.
Contact Benchling support to enable
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.
Eventbridge partner event source name now uses tenant ID [2026.3.DP.1]
When setting up a new partner event source in Benchling as part of an Eventbridge subscription, we now require using the tenant ID instead of the tenant name. This guarantees uniqueness and prevents issues from tenant renames. Note: this change does not impact existing Eventbridge subscriptions.
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.
Datetime parameters available for recipes and assays [2026.3.BP.1]
Create datetime parameters in unit operations. This is useful for process development teams that want to represent or capture: Material additions at a specific time Parameter shifts at a specific time Planned hold time Inoculation time
Lookup Studies [2026.3.BP.2]
Lookup tables/columns now allow you to look up the Studies associated with an object. For example a lookup table starting with an entity schema can be configured to include a column showing the Studies that any given entity is associated with.
Datetime parameters support minimum and maximum values [2026.3.BP.3]
Datetime data parameters in procedures now support minimum and maximum value constraints. Users can set date/time ranges when configuring procedure parameters, and the system validates that entered values fall within the specified range during procedure execution.
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.
Sequence limit warnings for charts [2026.3.PB.1]
Charts in PipeBio documents now display a warning when the number of sequences exceeds the supported limit for a given chart type, providing clearer feedback when working with large datasets.
Bug Fixes
Benchling AI
- Fixed an issue where Deep Research queries failed to end
- Fixed an issue where the Ask agent would fail when invoked by the Compose agent during agent switching.
- Fixed an issue where Compose would fail to fill plate or box tables for some users, displaying errors instead of populating the data.
- Fixed an issue where images generated during Deep Research would sometimes fail to display in the final answer. The system now correctly preserves generated image references when composing research reports.
- Fixed an issue where deep research would crash when searching and reading uploaded documents.
Notebook
- 24-hour time format now displays first hour times correctly
- Entry chip popover displays complete details
- Template collections can now be favorited
- Legacy workflow lookup field schema selection
- Fixed decimal field input in Mixture Prep Tables
- Prevent review submission without evaluation snapshot
- Mixture prep table submission audits now logged
- Fixed mixture prep table bug impacting pre-filling information after reordering columns
- Computed fields refresh on table load and block review
- Resolved issue with capturing rapid input into table cells
- Resolved template workspace tab issue showing entry icon when tag schema is assigned
- Subtemplate editing reliability improved
- Fix template collection creation in workspace
- Fixed a display issue showing spurious error indicators when previewing results tables prior to insertion
Molecular Biology
- Part link table mapping bug
- PCR product creation with existing primers
- Addgene accession ID imports disabled
- Fixed UniProt import bug
In Vivo
- Other task spec renaming error
- Donor ID randomization now works correctly
- Sampling plans display all study groups
- Sample plan validation causing error on study create
Workflows
- Bug fixed with worksheet stuck in review state after failed review submission
- Bug fixed with worksheet review with cross-step references and frozen chips
- Fixed file attachment bug on tasks
- Fixed a bug where the reviews service account was incorrectly attributed as the author of a worksheet review
Insights
- Resolve analyses that were unable to be opened due to incorrectly linked datasets
Schemas
- Updated schema configuration validation to catch conflicts with Fieldsets
- Reserved prefix error message clarified
- Archiving an already-archived object via V3 API incorrectly returns 500
Application Core
- Fixed bug loading printers label printing page
- Results tab performance improved for large entities
- Resolve intermittent errors with returning list of assay results from Benchling SDK
- Fixed an issue where an error was generated if the user navigated in the new config migration import UI before all the data was loaded.
- Configuration Migration -- Run schema imports with existing analysis templates
- V2 API -- results creation uses correct permissions
- Fixed issue where config migration import would be blocked if multiple mappings were found for the same object.
- Fixed an issue where a Benchling App could not be removed from a team in the Tenant Admin Console app tab. App could be removed from inside the team page, just not in the broader tenant page.
- Fixed an issue where User Admins weren't able to manage membership of some teams.
Compliance Controls
- Matrix scanner supports freeze-thaw tracked containers
- Text box data modifications save to audit log
- Folder exports support longer file names
- Fix for dropping file extension in s3 backups for entry attachments within structured tables
- HTML backups support descriptive filenames
BioProcess
- Locked procedures prevent metadata changes
PipeBio
- Fixed ENA import progress tracking
- Fixed annotation job hangs
- Fixed advanced filter grouping
- Fixed missing naming strategy option
- Fixed Benchling schema mapping crash
- Fixed structure prediction memory issue
- Fixed summary report error after clustering