Deprecations
Open official pageOpenAI deprecations
Source of record for the Assistants API deprecation notice, shutdown date, and recommended replacement.
OpenAI deprecation detail
Assistants API shutdown date: August 26, 2026.
This page answers the Assistants API shutdown date query directly. It keeps the shutdown date, replacement path, object-model changes, and migration checks in one source-linked view.
Why this matters now
The Assistants decision is no longer theoretical.
Use these points to frame whether the team should hold, migrate, or stop extending the current integration path.
Official sources
Check the OpenAI pages behind this migration brief.
This page uses a fixed source set so the migration brief stays auditable instead of drifting into secondhand advice.
Migration guide
Open official pageAssistants migration guide
Documents the object mapping from Assistants to Prompts, Threads to Conversations, Runs to Responses, and Run steps to Items.
Responses guide
Open official pageResponses migration guide
Explains the Responses API model, item-based request and output shape, and why new projects should prefer Responses.
API reference
Open official pageConversations API reference
Shows the Conversations resource and endpoints that replace Assistants Threads for stored conversation state.
Object mapping
The migration changes the objects your application reasons about.
The OpenAI migration guide changes the core units your application works with. These mappings are the minimum model shift a migration plan has to account for.
| Before | Now | What changes | Sources |
|---|---|---|---|
| Assistants | Prompts | Configuration moves to dashboard-managed prompts that are easier to version and update, rather than API-created assistant objects. | |
| Threads | Conversations | Conversation state now stores items, not just messages. OpenAI recommends moving new chats first and backfilling older threads only as needed. | |
| Runs | Responses | Execution becomes a Responses call that works from input items or a conversation object and returns typed output items. | |
| Run steps | Items | Tool calls, tool outputs, messages, and other execution artifacts move into an item-based model instead of run-step records. |
Migration checklist
Use the shutdown notice to drive a concrete sequence.
The checklist is ordered to reduce fresh debt first, then move conversation state and execution flow under control.
Step 1
Stop net-new Assistants adoption.
Treat Assistants as a migration target, not a platform choice. New chat paths should be planned around Responses plus Conversations instead of extending the deprecated surface.
Sources
Step 2
Recreate important assistants as prompts.
Identify the assistant objects that matter, recreate them as prompts, and store stable prompt identifiers so application code can refer to versioned behavior instead of assistant objects.
Sources
Step 3
Move new chats first, then backfill only where needed.
OpenAI will not provide an automated Threads-to-Conversations migration tool. Start new user chats on Conversations and backfill older Threads only where history continuity is worth the effort.
Step 4
Update request and output handling for Responses.
The execution path changes from Runs and run steps to Responses and Items. Your integration needs to read and validate the new response object and item model before cutover.
Verify before cutover
Do not treat the route swap as the end of the migration.
These checks are the minimum confidence gates before any production traffic is moved off Assistants.
Prompt recreation and versioning
Confirm each migrated assistant has a prompt owner, a versioning plan, and a stable prompt ID or exported spec that the application can reference.
Sources
Conversation migration strategy
Decide what starts fresh in Conversations, which legacy Threads need backfill, and how metadata or continuity should be preserved across the move.
Tool and runtime assumptions
Validate that your tool loops, stored conversation state, and response handling still behave the way production expects under the Responses and Conversations model.
End-to-end acceptance before switching traffic
Run staging flows that cover common conversations, tool usage, retries, and failure handling before moving production traffic off Assistants.
Sources
Continue the site
Keep moving through the decision from here.
Use the groups below to move laterally through the decision, not back out into another doc hunt.
Related pages
Stay in the same decision neighborhood instead of backing out to search.
OpenAI deprecations tracker
Provider-specific hub for OpenAI shutdown dates, migration briefs, and replacement paths.
Open pageRealtime API beta shutdown
Shutdown-date brief for the Realtime beta interface and the current GA migration path.
Open pageLegacy GPT snapshots
Migration brief for dated legacy GPT snapshots and older preview aliases with a fixed cutoff.
Open pageCompare pages
Open the pages that turn this topic into a side-by-side decision.
Replacement pages
Use the likely substitutes, migration targets, or fallback choices as the next click.
Source category pages
Trace the source families behind this page instead of opening random docs in isolation.