OpenAI has doubled usage quotas for several paid tiers, directly expanding Work and Codex limits on its Pro 100 and Pro 500 plans. On the Pro 200 plan, the company reversed a previously scheduled reduction in capacity following direct pushback from users, while quotas for the entry-level Plus tier remain unchanged. New subscriptions incorporate the updated limits immediately, and existing subscribers receive the expanded allocations automatically.
According to OpenAI, doubling a quota does not translate to a linear doubling of completed tasks, as token consumption fluctuates based on the underlying model, task complexity, and user settings. The expanded allocations for the newly introduced "Dots" interface remain active, though usage for these components is monitored independently from standard workloads. This dynamic directly impacts the unit economics of enterprise deployments, where shifting infrastructure costs and unpredictable token utilization complicate predictable budgeting for heavy development tasks.
As OpenAI stated, the company intends to maintain its deployment of model updates and infrastructure optimizations to drive usable capacity upward, managing the underlying compute overhead required for expanded limits. This operational adjustment follows the recent rollout of the 6.1 model generation, comprising Terra and Sol, which deliver higher operational efficiency while matching or exceeding the benchmark performance of their predecessors. The capacity adjustments arrive as Anthropic prepares for its public offering, intensifying the direct competition between both providers for enterprise software budgets and B2B market share as firms weigh infrastructure expenses against model performance.
Editorial analysis
Corporate AI spending hit a ceiling in July, but by early October OpenAI was expanding capacity on its premium tiers under pressure from Anthropic's planned IPO. Enterprise consumption jumped about 50 percent by late September, pushing the provider to absorb higher Work and Codex limits on Pro 100 and Pro 500 subscriptions. The expanded allowance does not ensure an equivalent volume of completed work because token consumption shifts with every task.
Our stories: OpenAI and Anthropic Price War Cuts Enterprise AI Costs (October 2).