Thinking Machine Labs Inkling model scheduled for retirement on Databricks

The Thinking Machine Labs Inkling language model is scheduled for removal from Databricks. Starting October 30, 2026, it will no longer function on Foundation Model APIs, and workloads using it must be migrated to newer options.

Use cases

Imagine you have production applications consuming Inkling for language tasks and need to shift your pipelines to supported models before the endpoint stops responding. You can:

• Identify which endpoints and users still depend on the deprecated model.
• Update your applications to use the specified replacements.

How to use

Run a SQL query in Databricks to find active workloads using the old model and discover their respective owners.

SELECT
   eu.requester,
   se.endpoint_name,
   se.entity_name,
   COUNT(*) AS request_count,
   SUM(eu.input_token_count) AS total_input_tokens,
   SUM(eu.output_token_count) AS total_output_tokens,
   MIN(eu.request_time) AS first_request,
   MAX(eu.request_time) AS last_request
 FROM system.serving.endpoint_usage eu
 JOIN system.serving.served_entities se
   ON eu.served_entity_id = se.served_entity_id
 WHERE LOWER(se.entity_name) LIKE '%thinking-machine-labs-inkling%'
 GROUP BY eu.requester, se.endpoint_name, se.entity_name
 ORDER BY request_count DESC

What happens in this code:

1. Joins endpoint usage data with the served entities table in the system catalog.
2. Filters by the name of the model slated for retirement.
3. Groups requests by requesting user and endpoint name to audit the usage volume.

Limitations and requirements

The Inkling model is currently deprecated and will be fully retired on October 30, 2026. Databricks recommends migrating to GLM 5.3 or Kimi K3 as replacement models.