Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 4122c7f998
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| "context": 128000, | ||
| "output": 262144 |
There was a problem hiding this comment.
This update reduces poe/novita/kimi-k2.5 context from 256000 to 128000 while leaving output at 262144, so the advertised max completion now exceeds the total context window. Downstream consumers that use these limits for token budgeting (UI display, validation, or truncation logic) will be working from an impossible configuration and can mislead users or over-allocate tokens. Please align limit.output to be <= limit.context (or restore the prior context value if 128000 is not intended).
Useful? React with 👍 / 👎.
|
There are a bunch of changes to other models here as well, is that intentional? |
3 participants
Opus 4.7 is out - this teaches goose about it - context size etc - just routine update.
This description was written by hand, by mic, old school.