I use spark somewhat frequently but it has disappeared from my model selection in the VSCode Plugin with the latest update with the new reasoning selector.
I have 100% of spark usage remaining
I found about half a dozen similar posts over the past few days but most ended in not being on pro, looking in the wrong place, or some other general misunderstanding. There was one post that said it wasn’t available with ChatGPT Pro subscriptions, but I have been using it so I am hoping that it isn’t something they have taken away.
As much as this sounds like a bug, I did some Q&A with ChatGPT and found the following:
OpenAI’s current documentation lists GPT-5.3-Codex-Spark as available to ChatGPT Pro users and shows it as available through the Codex CLI and IDE extension.
Have you tried starting Codex from the command line and explicitly specifying the model?
codex --model gpt-5.3-codex-spark
That might help narrow down where the problem is. If Spark works from the CLI but is missing from the VS Code model selector, that would point more toward a VS Code extension/model-selector issue than an account-level availability problem.
The problem appears to be one of the following:
an OpenAI documentation gap or outdated Help Center article, where the relevant information at https://help.openai.com/en has not yet been updated to reflect the current availability of GPT-5.3-Codex-Spark and the Codex surfaces that support it;
a VS Code extension/model-selector regression, especially since Spark was available before the extension update;
an account entitlement or rollout synchronization issue, where the Spark quota is visible but the model is not being exposed to that client;
a local Codex configuration, authentication, or client-state issue;
or another Codex bug related to model availability.
If Spark is also unavailable when explicitly selected from the CLI, that would make an account/availability issue more likely. In that case, you could also consider opening a ticket with OpenAI Support.
It still looks sufficiently bug-like to me, especially since you were using Spark before the update, so I am leaving this in the Codex - Bugs subcategory.
