Pro Lite account receives 272K context profile for gpt-5.6-sol instead of bundled 372K
What version of Codex CLI is running?
OpenAI Codex (v0.144.4)
What subscription do you have?
(Pro Lite)
Which model were you using?
gpt-5.6-sol (reasoning medium, summaries auto)
What platform is your computer?
_No response_
What terminal emulator and version are you using (if applicable)?
_No response_
Codex doctor report
What issue are you seeing?
Description
After upgrading from ChatGPT Plus to Pro Lite, Codex CLI initially continued to identify my account as Plus. Submitting feedback caused Codex to sign me out, and after reauthentication the account is now correctly shown as Pro Lite and the usage limits increased.
However, the context window for gpt-5.6-sol remains limited to 272,000 raw tokens / approximately 258,400 effective tokens.
Environment
- Codex CLI:
0.144.3 - Platform: Ubuntu Linux x86_64
- Subscription: Pro Lite
- Model:
gpt-5.6-sol - Reasoning effort: Medium
- Fresh thread context shown by
/status:258K - Local
model_context_windowoverride: none - Local custom model catalog: none
Model catalog results
Bundled catalog:
gpt-5.6-sol
context_window: 372000
max_context_window: 372000
Current/effective catalog:
gpt-5.6-sol
context_window: 272000
max_context_window: 272000
effective_context_window_percent: 95
This results in:
272000 × 95% = 258400
which matches the fresh-thread context shown by /status.
Question
Is the 272,000-token Sol context window intentional for Pro Lite subscriptions, or is my account receiving a legacy or reduced model profile?
The subscription entitlement appears to have refreshed correctly because Codex now identifies the account as Pro Lite and the usage allowance increased, but the remote/current Sol model catalog still overrides the bundled 372,000-token profile with a 272,000-token profile.
Uploaded diagnostic thread
019f60cf-1ff0-7db2-b56c-53b387c810a3
What steps can reproduce the bug?
Uploaded thread: 019f60cf-1ff0-7db2-b56c-53b387c810a3
What is the expected behavior?
_No response_
Additional information
_No response_
This issue has 1 comment on GitHub. Read the full discussion on GitHub ↗