As of November 12, 2026, OpenAI marks gpt-5-chat-latest shut down. Its model page says it pointed to the GPT-5 snapshot previously used in ChatGPT; it directs readers seeking the latest Chat model to the separate chat-latest entry. OpenAI describes chat-latest as a regularly updated pointer to the latest Instant model used in ChatGPT—but does not describe it as an equivalent replacement for gpt-5-chat-latest. (GPT-5 Chat model page; Chat Latest model page)
What does “GPT-5 chat model” mean in the API?
The exact legacy API ID is gpt-5-chat-latest, and its official model entry is marked shut down. That entry identifies it as the GPT-5 snapshot previously used in ChatGPT. The current Chat Latest entry has a different ID—chat-latest—and points to the latest Instant model used in ChatGPT. OpenAI says the snapshot behind chat-latest is regularly updated. The documentation establishes these different identities and statuses, not a migration or equivalence between the two IDs. (GPT-5 Chat model page; Chat Latest model page)
For a new integration, don’t use the shut-down GPT-5 ID. If you specifically need the current ChatGPT Instant pointer, chat-latest is the documented ID. Since it changes behind the alias, evaluate it with representative requests rather than assuming behavior will remain identical over time. (Chat Latest model page)
What workloads is the current chat model suited to?
OpenAI documents chat-latest with text input and output and image input. For the Responses API, its listed supported tools include web search, file search, image generation, Code Interpreter, and MCP. The same catalog lists audio, video, and computer use as unsupported. If your workflow requires function calling or structured outputs, verify support for the exact model ID in the current documentation before choosing it. Match the documented features to your workflow, then test answer quality and tool behavior on representative tasks; the feature list alone does not establish how well the model will meet your application’s quality or latency targets. (Chat Latest model page)
OpenAI recommends GPT-6 Astra for production API usage on the Chat Latest page. That is a recommendation, not a documented guarantee of fixed behavior or reproducibility. Compare models against your workload and validate again if you rely on a rolling alias. (Chat Latest model page; GPT-6 Astra model page)
For document retrieval, the catalog lists file search for chat-latest through the Responses API. OpenAI’s file-search guide says you need a vector store with files uploaded before using the tool. Verify the tool setup and retrieval results with your own documents. (Chat Latest model page; File search guide)
Current price and limits for chat-latest
OpenAI lists $5.00 per 1 million input tokens, $0.50 per 1 million cached input tokens, and $30.00 per 1 million output tokens. Its model page lists a 400,000-token context window and a 128,000-token maximum output. Check the live pricing page and your account’s rate limits when planning a deployment. (OpenAI API pricing; Chat Latest model page)
A minimal API check
Install the official Python SDK and configure OPENAI_API_KEY in your environment; the SDK’s OpenAI() client reads that variable by default. This example adapts the SDK’s documented Responses API pattern to the chat-latest model ID listed in OpenAI’s model catalog. It is an example to try, not a claim that the request was tested here. (OpenAI Python SDK; Chat Latest model page)
from openai import OpenAI
client = OpenAI()
response = client.responses.create(
model="chat-latest",
input="In one sentence, explain what a rolling model alias is."
)
print(response.output_text)
The model catalog lists the Responses API as a supported endpoint. Confirm that the model is available to your API account and check the response, errors, tool behavior, and usage in your own environment before deployment. If you use file search, prepare the vector store and uploaded files as described in OpenAI’s guide. (Chat Latest model page; File search guide)
Bottom line
gpt-5-chat-latest is marked shut down. chat-latest is the separately documented, regularly updated Instant-model pointer; OpenAI does not document it as a direct replacement or equivalent for the shut-down GPT-5 entry. For production API use, OpenAI recommends evaluating GPT-6 Astra. Choose based on documented tool and modality requirements, test with representative tasks, and account for the changing snapshot behind chat-latest. (GPT-5 Chat model page; Chat Latest model page; GPT-6 Astra model page)
Sources
Chat Latest Model | OpenAI API