OpenAI has reinstated a rolling five-hour rate limit on Codex and ChatGPT Work for ChatGPT Plus subscribers. The change ends a temporary multi-week period during which the short-term window was suspended and users were constrained only by overall weekly quota ceilings.
OpenAI engineering lead Thibault Sottiaux confirmed the reinstatement, stating that re-enforcing the five-hour window is necessary to smooth peak compute load across OpenAI's inference fleet as developer adoption of complex multi-agent coding sessions and large-context reasoning models expands.
Balancing Agentic Workloads and Fleet Capacity
During the summer trial period, OpenAI removed the five-hour cap to monitor user behavior and assess infrastructure limits following major active user growth across unified developer environments. However, long-running agentic tasks, large codebase indexing, and frequent subagent spawns created spiky GPU utilization patterns.
Under the restored policy:
- Dual-layer quotas: ChatGPT Plus accounts are subject to both a rolling five-hour limit and a secondary weekly volume cap.
- Over-limit behavior: Users who hit the five-hour window can wait for the rolling reset or purchase supplemental compute credits.
- Token optimizations: OpenAI engineers implemented targeted token reduction improvements across reasoning models, including GPT-5.6 Sol, to reduce per-session context overhead.
The restored policy reflects the ongoing tension between offering unrestricted interactive developer workflows and managing cluster contention across multi-tenant frontier model serving infrastructures.
