<!--IssueSummary start--> <details> <summary> Everyone can contribute. [Help move this issue forward](https://handbook.gitlab.com/handbook/marketing/developer-relations/contributor-success/community-contributors-workflows/#contributor-links) while earning points, leveling up and collecting rewards. </summary> - [Collaborate/take over this issue](https://contributors.gitlab.com/manage-issue?action=work&projectId=278964&issueIid=547122) </details> <!--IssueSummary end--> There are currently following endpoints on AI GW side used actively for code suggestion requests: * /v2/code/completions and /v2/completions - used for code completion requests (v2 uses different format request params) * /v3/code/completions - used for code generation requests which don't support SSE streaming * /v4/code/suggestions - used for code generation requests which support SSE streaming (API for v3 vs v4 endpoints is same for request's params, but response format is different) Both v3 and v4 endpoint support both completion and generation requests (the request type is passed as another request param). A downside is that we have to maintain multiple versions of endpoints which adds extra maintenance overhead (which we will have to maintain another [2 major versions](https://handbook.gitlab.com/handbook/engineering/architecture/design-documents/ai_gateway/#basic-stable-api-for-the-ai-gateway). It would be better if we could switch to newest v4 endpoint both for completions and generations and deprecate v2 and v3 endpoints. Related older issues: * v4 endpoint https://gitlab.com/gitlab-org/modelops/applied-ml/code-suggestions/ai-assist/-/merge_requests/1637 * v3 migration https://gitlab.com/gitlab-org/gitlab/-/issues/477891#note_2059060535