-1.png?w=1024)
OpenAI is giving eligible paid ChatGPT users more access to GPT-5.6 Sol after reporting stronger-than-expected early demand for the reasoning model.
The company temporarily removed the model’s five-hour usage window for eligible paid subscribers and reset existing usage quotas. OpenAI said the model’s improved efficiency made the temporary access expansion possible.
The move follows what OpenAI described as an unexpectedly strong wave of usage during GPT-5.6 Sol’s first 48 hours of availability, particularly across coding and AI agent workflows. The expansion gives power users more uninterrupted time with the reasoning model, although OpenAI has not said how long the relaxed limits will remain in place.
Why OpenAI eased GPT-5.6 Sol’s limits
OpenAI introduced GPT-5.6 as a three-model family designed for different workloads.
- Sol serves as the highest-capability reasoning model.
- Terra is positioned as a balanced, middle-tier model.
- Luna is the most cost-efficient tier.
The release, however, saw stronger-than-expected demand. In an X post, Thibault (Tibo) Sottiaux, head of engineering at Codex, revealed that within 48 hours of its release, Codex alongside ChatGPT Work reached 6 million active users.
.png)
OpenAI engineering executive Thibault Sottiaux announces expanded GPT-5.6 Sol access. Image: X
The previous five-hour quota window has been lifted for eligible ChatGPT Plus, Pro, and Business subscribers. At the same time, users who had already exhausted their allowance have had their quotas reset.
The company did not indicate whether the relaxed limits will become permanent, leaving open the possibility that usage restrictions could be adjusted again as demand stabilizes.
Efficiency is increasingly becoming as important as intelligence
In his X post, Sottiaux added that the GPT-5.6 Sol update would result in less token usage per completed task.
That update reflects a growing reality across the AI industry: improving model capability is no longer enough on its own. As models improve their reasoning and agentic capabilities, companies are under increasing pressure to make those systems more efficient without sacrificing performance.
The shift became particularly visible as Chinese AI developers released models that rival leading Western systems while emphasizing lower training and inference costs. This has become especially important amid rising demand for costly AI computing capacity, suggesting that winning the race now depends as much on aggressive efficiency as on model intelligence. OpenAI now appears to be working toward that goal with this new move.
What this means for ChatGPT users
The immediate benefit is straightforward: longer uninterrupted access to OpenAI’s most capable reasoning model. Developers, researchers, and other power users can spend more time coding, debugging, and tackling complex reasoning tasks without having to return to the previous usage window quickly. At the same time, those who have already exhausted their allowance receive a fresh allocation.
The changes may also benefit organizations experimenting with AI agents and software development workflows. Fewer interruptions mean teams can spend less time working around usage constraints and more time evaluating how frontier reasoning models fit into everyday work.
More News: OpenAI's new ChatGPT Work agent can handle complex, multi-step tasks across apps, files, and workflows, marking the company's biggest push yet into enterprise AI assistants.