[9.2] [ML][Inference Endpoints] Anthropic endpoint creation: ensure max tokens parameter is passed as expected (#241212) (#241343)

# Backport

This will backport the following commits from `main` to `9.2`:
- [[ML][Inference Endpoints] Anthropic endpoint creation: ensure max
tokens parameter is passed as expected
(#241212)](#241212)

<!--- Backport version: 9.6.6 -->

### Questions ?
Please refer to the [Backport tool
documentation](https://github.com/sorenlouv/backport)

<!--BACKPORT [{"author":{"name":"Melissa
Alvarez","email":"[email protected]"},"sourceCommit":{"committedDate":"2025-10-30T17:36:58Z","message":"[ML][Inference
Endpoints] Anthropic endpoint creation: ensure max tokens parameter is
passed as expected (#241212)\n\n## Summary\n\nRelated to this
[issue](https://github.com/elastic/kibana/issues/241142)\nand this
[fix](https://github.com/elastic/kibana/pull/241188).\n\nThis PR:\n-
updates the inference creation endpoint to ensure max_tokens are
sent\ncorrectly for Anthropic\n- ensures that max_tokens is added back
into the providerConfig when\nviewing the endpoint so that it shows up
correctly in the form\n\nThis is a temporary workaround for anthropic
max_tokens handling until\nthe services endpoint is updated to reflect
the correct structure.\nAnthropic is unique in that it requires
max_tokens to be sent as part of\nthe task_settings instead of the usual
service_settings.\nUntil the services endpoint is updated to reflect
that, there is no way\nfor the form UI to know where to put max_tokens.
This can be removed\nonce that update is made.\n\n\n###
Checklist\n\nCheck the PR satisfies following conditions. \n\nReviewers
should verify this PR satisfies this list as well.\n\n- [ ] Any text
added follows [EUI's
writing\nguidelines](https://elastic.github.io/eui/#/guidelines/writing),
uses\nsentence case text and includes
[i18n\nsupport](https://github.com/elastic/kibana/blob/main/src/platform/packages/shared/kbn-i18n/README.md)\n-
[
]\n[Documentation](https://www.elastic.co/guide/en/kibana/master/development-documentation.html)\nwas
added for features that require explanation or tutorials\n- [ ] [Unit or
functional\ntests](https://www.elastic.co/guide/en/kibana/master/development-tests.html)\nwere
updated or added to match the most common scenarios\n- [ ] If a plugin
configuration key changed, check if it needs to be\nallowlisted in the
cloud and added to the
[docker\nlist](https://github.com/elastic/kibana/blob/main/src/dev/build/tasks/os_packages/docker_generator/resources/base/bin/kibana-docker)\n-
[ ] This was checked for breaking HTTP API changes, and any
breaking\nchanges have been approved by the breaking-change committee.
The\n`release_note:breaking` label should be applied in these
situations.\n- [ ] [Flaky
Test\nRunner](https://ci-stats.kibana.dev/trigger_flaky_test_runner/1)
was\nused on any tests changed\n- [ ] The PR description includes the
appropriate Release Notes section,\nand the correct `release_note:*`
label is applied per
the\n[guidelines](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process)\n-
[ ] Review the
[backport\nguidelines](https://docs.google.com/document/d/1VyN5k91e5OVumlc0Gb9RPa3h1ewuPE705nRtioPiTvY/edit?usp=sharing)\nand
apply applicable `backport:*`
labels.","sha":"847f9de184d2918f261148ee62350e22bf7e079b","branchLabelMapping":{"^v9.3.0$":"main","^v(\\d+).(\\d+).\\d+$":"$1.$2"}},"sourcePullRequest":{"labels":["release_note:fix",":ml","backport:version","Feature:Inference
UI","v9.3.0","v8.19.7","v9.1.7","v9.2.1"],"title":"[ML][Inference
Endpoints] Anthropic endpoint creation: ensure max tokens parameter is
passed as
expected","number":241212,"url":"https://github.com/elastic/kibana/pull/241212","mergeCommit":{"message":"[ML][Inference
Endpoints] Anthropic endpoint creation: ensure max tokens parameter is
passed as expected (#241212)\n\n## Summary\n\nRelated to this
[issue](https://github.com/elastic/kibana/issues/241142)\nand this
[fix](https://github.com/elastic/kibana/pull/241188).\n\nThis PR:\n-
updates the inference creation endpoint to ensure max_tokens are
sent\ncorrectly for Anthropic\n- ensures that max_tokens is added back
into the providerConfig when\nviewing the endpoint so that it shows up
correctly in the form\n\nThis is a temporary workaround for anthropic
max_tokens handling until\nthe services endpoint is updated to reflect
the correct structure.\nAnthropic is unique in that it requires
max_tokens to be sent as part of\nthe task_settings instead of the usual
service_settings.\nUntil the services endpoint is updated to reflect
that, there is no way\nfor the form UI to know where to put max_tokens.
This can be removed\nonce that update is made.\n\n\n###
Checklist\n\nCheck the PR satisfies following conditions. \n\nReviewers
should verify this PR satisfies this list as well.\n\n- [ ] Any text
added follows [EUI's
writing\nguidelines](https://elastic.github.io/eui/#/guidelines/writing),
uses\nsentence case text and includes
[i18n\nsupport](https://github.com/elastic/kibana/blob/main/src/platform/packages/shared/kbn-i18n/README.md)\n-
[
]\n[Documentation](https://www.elastic.co/guide/en/kibana/master/development-documentation.html)\nwas
added for features that require explanation or tutorials\n- [ ] [Unit or
functional\ntests](https://www.elastic.co/guide/en/kibana/master/development-tests.html)\nwere
updated or added to match the most common scenarios\n- [ ] If a plugin
configuration key changed, check if it needs to be\nallowlisted in the
cloud and added to the
[docker\nlist](https://github.com/elastic/kibana/blob/main/src/dev/build/tasks/os_packages/docker_generator/resources/base/bin/kibana-docker)\n-
[ ] This was checked for breaking HTTP API changes, and any
breaking\nchanges have been approved by the breaking-change committee.
The\n`release_note:breaking` label should be applied in these
situations.\n- [ ] [Flaky
Test\nRunner](https://ci-stats.kibana.dev/trigger_flaky_test_runner/1)
was\nused on any tests changed\n- [ ] The PR description includes the
appropriate Release Notes section,\nand the correct `release_note:*`
label is applied per
the\n[guidelines](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process)\n-
[ ] Review the
[backport\nguidelines](https://docs.google.com/document/d/1VyN5k91e5OVumlc0Gb9RPa3h1ewuPE705nRtioPiTvY/edit?usp=sharing)\nand
apply applicable `backport:*`
labels.","sha":"847f9de184d2918f261148ee62350e22bf7e079b"}},"sourceBranch":"main","suggestedTargetBranches":["8.19","9.1","9.2"],"targetPullRequestStates":[{"branch":"main","label":"v9.3.0","branchLabelMappingKey":"^v9.3.0$","isSourceBranch":true,"state":"MERGED","url":"https://github.com/elastic/kibana/pull/241212","number":241212,"mergeCommit":{"message":"[ML][Inference
Endpoints] Anthropic endpoint creation: ensure max tokens parameter is
passed as expected (#241212)\n\n## Summary\n\nRelated to this
[issue](https://github.com/elastic/kibana/issues/241142)\nand this
[fix](https://github.com/elastic/kibana/pull/241188).\n\nThis PR:\n-
updates the inference creation endpoint to ensure max_tokens are
sent\ncorrectly for Anthropic\n- ensures that max_tokens is added back
into the providerConfig when\nviewing the endpoint so that it shows up
correctly in the form\n\nThis is a temporary workaround for anthropic
max_tokens handling until\nthe services endpoint is updated to reflect
the correct structure.\nAnthropic is unique in that it requires
max_tokens to be sent as part of\nthe task_settings instead of the usual
service_settings.\nUntil the services endpoint is updated to reflect
that, there is no way\nfor the form UI to know where to put max_tokens.
This can be removed\nonce that update is made.\n\n\n###
Checklist\n\nCheck the PR satisfies following conditions. \n\nReviewers
should verify this PR satisfies this list as well.\n\n- [ ] Any text
added follows [EUI's
writing\nguidelines](https://elastic.github.io/eui/#/guidelines/writing),
uses\nsentence case text and includes
[i18n\nsupport](https://github.com/elastic/kibana/blob/main/src/platform/packages/shared/kbn-i18n/README.md)\n-
[
]\n[Documentation](https://www.elastic.co/guide/en/kibana/master/development-documentation.html)\nwas
added for features that require explanation or tutorials\n- [ ] [Unit or
functional\ntests](https://www.elastic.co/guide/en/kibana/master/development-tests.html)\nwere
updated or added to match the most common scenarios\n- [ ] If a plugin
configuration key changed, check if it needs to be\nallowlisted in the
cloud and added to the
[docker\nlist](https://github.com/elastic/kibana/blob/main/src/dev/build/tasks/os_packages/docker_generator/resources/base/bin/kibana-docker)\n-
[ ] This was checked for breaking HTTP API changes, and any
breaking\nchanges have been approved by the breaking-change committee.
The\n`release_note:breaking` label should be applied in these
situations.\n- [ ] [Flaky
Test\nRunner](https://ci-stats.kibana.dev/trigger_flaky_test_runner/1)
was\nused on any tests changed\n- [ ] The PR description includes the
appropriate Release Notes section,\nand the correct `release_note:*`
label is applied per
the\n[guidelines](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process)\n-
[ ] Review the
[backport\nguidelines](https://docs.google.com/document/d/1VyN5k91e5OVumlc0Gb9RPa3h1ewuPE705nRtioPiTvY/edit?usp=sharing)\nand
apply applicable `backport:*`
labels.","sha":"847f9de184d2918f261148ee62350e22bf7e079b"}},{"branch":"8.19","label":"v8.19.7","branchLabelMappingKey":"^v(\\d+).(\\d+).\\d+$","isSourceBranch":false,"state":"NOT_CREATED"},{"branch":"9.1","label":"v9.1.7","branchLabelMappingKey":"^v(\\d+).(\\d+).\\d+$","isSourceBranch":false,"state":"NOT_CREATED"},{"branch":"9.2","label":"v9.2.1","branchLabelMappingKey":"^v(\\d+).(\\d+).\\d+$","isSourceBranch":false,"state":"NOT_CREATED"}]}]
BACKPORT-->

Co-authored-by: Melissa Alvarez <[email protected]>

Author: kibanamachine

x-pack/platform/plugins/shared/inference_endpoint/server/routes/index.ts

@@ -106,6 +106,18 @@ export const getInferenceServicesRoute = (
 
           const { config, secrets } = request.body;
 
+          // NOTE: This is a temporary workaround for anthropic max_tokens handling until the services endpoint is updated to reflect the correct structure.
+          // Anthropic is unique in that it requires max_tokens to be sent as part of the task_settings instead of the usual service_settings.
+          // Until the services endpoint is updated to reflect that, there is no way for the form UI to know where to put max_tokens. This can be removed once that update is made.
+          let taskSettings;
+          if (config?.provider === 'anthropic' && config?.providerConfig?.max_tokens) {
+            taskSettings = {
+              max_tokens: config.providerConfig.max_tokens,
+            };
+            // This field is unknown to the anthropic service config, so we remove it
+            delete config.providerConfig.max_tokens;
+          }
+
           const serviceSettings = {
             ...unflattenObject(config?.providerConfig ?? {}),
             ...unflattenObject(secrets?.providerSecrets ?? {}),
@@ -117,6 +129,7 @@ export const getInferenceServicesRoute = (
             inference_config: {
               service: config?.provider,
               service_settings: serviceSettings,
+              ...(taskSettings ? { task_settings: taskSettings } : {}),
             },
           });
 

x-pack/solutions/search/plugins/search_inference_endpoints/public/components/edit_inference_endpoints/edit_inference_flyout.tsx

@@ -10,6 +10,7 @@ import React, { useCallback } from 'react';
 import type { InferenceEndpoint } from '@kbn/inference-endpoint-ui-common';
 import { flattenObject } from '@kbn/object-utils';
 import type { InferenceInferenceEndpointInfo } from '@elastic/elasticsearch/lib/api/types';
+import { ServiceProviderKeys } from '@kbn/inference-endpoint-ui-common';
 import { useKibana } from '../../hooks/use_kibana';
 import { useQueryInferenceEndpoints } from '../../hooks/use_inference_endpoints';
 
@@ -38,7 +39,15 @@ export const EditInferenceFlyout: React.FC<EditInterfaceFlyoutProps> = ({
       inferenceId: selectedInferenceEndpoint.inference_id,
       taskType: selectedInferenceEndpoint.task_type,
       provider: selectedInferenceEndpoint.service,
-      providerConfig: flattenObject(selectedInferenceEndpoint.service_settings),
+      providerConfig: {
+        ...flattenObject(selectedInferenceEndpoint.service_settings),
+        // NOTE: The below is a workaround for anthropic max_tokens handling.
+        // Anthropic is unique in that it requires max_tokens to be stored as part of the task_settings instead of the usual service_settings - which we populate the providerConfig from.
+        ...(selectedInferenceEndpoint.task_settings?.max_tokens &&
+        selectedInferenceEndpoint.service === ServiceProviderKeys.anthropic
+          ? { max_tokens: selectedInferenceEndpoint.task_settings?.max_tokens }
+          : {}),
+      },
     },
     secrets: {
       providerSecrets: {},