OpenAI lifts GPT-5.6 Sol limits as Codex hits 8 million users

OpenAI's decision to ease GPT-5.6 Sol usage caps comes as its coding platform Codex crosses 8 million active users, as enterprise demand surges.

OpenAI temporarily removed the five-hour usage window for GPT-5.6 Sol across Plus, Business, and Pro plans after Codex and ChatGPT Work usage surged past 8 million active users, the company said July 12.

"The last 48 hours of Codex and ChatGPT Work have been intense," Tibo Sottiaux, product lead at OpenAI, said in a post on X. The company is rolling out inference optimizations that will make Sol more efficient, reducing per-task consumption by about 10 percent.

The change eliminates the rolling five-hour restriction that previously forced users to pause work, though separate weekly caps remain in place. OpenAI also granted every Codex and ChatGPT Work account a one-time banked reset that can be saved and applied to replenish weekly usage. The company temporarily reduced a product context setting to 272,000 tokens from 372,000 because the larger setting charged more quota than intended, with plans to restore it after further work.

The moves reflect OpenAI's push to convert surging trial usage into recurring subscription revenue. Codex's 8 million users — up from 7 million just days earlier — represent a rapidly expanding base that could drive higher retention and reduce churn, key metrics for any AI platform seeking to justify its valuation.

Competitors Hold the Line on Caps

Google's Gemini Apps retain both five-hour and weekly compute-based limits, while Anthropic extended free access to Claude Fable 5 on paid plans through July 19 and kept Claude Code's weekly limits 50 percent above normal. None of the three companies has permanently abandoned usage caps, but OpenAI's temporary relaxation gives it a short-term edge in user experience at a moment when enterprise teams are evaluating which platform to standardize on.

The timing is strategic. GPT-5.6 Sol, released July 9, offers a 1,050,000-token context window — the largest among frontier models — and supports multi-agent workflows that can parallelize complex coding tasks. Those capabilities consume significant compute, which is why OpenAI initially imposed the five-hour window. By temporarily removing it while improving inference efficiency, the company is betting that better user experience will drive habit formation faster than strict quota enforcement would.

The Revenue Calculus for Investors

For investors, the math is straightforward: 8 million active users across Codex and ChatGPT Work, each on a paid plan ranging from $20 per month for Plus to $200 for Pro, implies an annualized revenue run rate in the billions. Every percentage point improvement in retention compounds significantly at that scale. Inference optimizations that reduce per-task token consumption also improve unit economics, widening margins on each subscription dollar.

OpenAI did not disclose how long the relaxed limits will remain in place or whether the change will become permanent. Sottiaux said the company is "rolling out changes that will make GPT-5.6 Sol more efficient across the board," suggesting the temporary measure could become the new normal if infrastructure capacity keeps pace with demand.

This article is for informational purposes only and does not constitute investment advice.