Client or integration
Codex CLI
Area
CLI
Summary
When working with Codex authentication, a secondary Pro account (with over 50% quota remaining) was added after the primary Plus account's token quota was already fully depleted. However, instead of smoothly routing requests through the secondary account, execution frequently halts mid-task or fails with an overuse warning exclamation icon.
Expected:
Since a secondary Pro account with available tokens was registered, the tool should seamlessly utilize the secondary account's quota without freezing or requiring manual retries when the primary account remains exhausted.
Actual:
Execution frequently stalls during sessions, displaying a limit-reached exclamation mark. In most cases, operations freeze entirely, requiring manual user intervention (e.g., repeatedly instructing it to retry) to resume progress.
Reproduction
- Use a primary Codex account (Plus tier) whose token quota has already been completely exhausted.
- Subsequently add a secondary Codex authentication account (Pro tier with >50% token quota remaining) via ocx gui.
- Initiate a new session or multi-step prompt that consumes model tokens.
- Observe execution halting during active tasks, accompanied by an overuse/limit status indicator (exclamation mark icon).
- Note that the process does not automatically proceed with the Pro account, but instead freezes until manually prompted to retry.
Observed Conditions:
- Occurs repeatedly across multiple sessions while operating under this account configuration.
- Manifests as an unhandled execution stall, requiring explicit manual input ("retry") to continue work.
Version
2.7.41
Operating system
macOS 26.5.2(25F84)
Provider and model
GPT-5.6(Luna, Terra, Sol)
Logs or error output
Screenshots and supporting files
Redacted configuration
Checks
Client or integration
Codex CLI
Area
CLI
Summary
When working with Codex authentication, a secondary Pro account (with over 50% quota remaining) was added after the primary Plus account's token quota was already fully depleted. However, instead of smoothly routing requests through the secondary account, execution frequently halts mid-task or fails with an overuse warning exclamation icon.
Expected:
Since a secondary Pro account with available tokens was registered, the tool should seamlessly utilize the secondary account's quota without freezing or requiring manual retries when the primary account remains exhausted.
Actual:
Execution frequently stalls during sessions, displaying a limit-reached exclamation mark. In most cases, operations freeze entirely, requiring manual user intervention (e.g., repeatedly instructing it to retry) to resume progress.
Reproduction
Observed Conditions:
Version
2.7.41
Operating system
macOS 26.5.2(25F84)
Provider and model
GPT-5.6(Luna, Terra, Sol)
Logs or error output
Screenshots and supporting files
Redacted configuration
Checks