OpenAI has flipped a short-lived switch on GPT-5.6 Sol: the tight five-hour throttle on heavy users has been loosened and current usage counters were reset. If you’ve been banging away at agentic workflows or furious coding sessions, this is the small mercy you were waiting for.

What changed (short version)

Here’s the gist without the corporate-speak: OpenAI temporarily removed the five-hour window cap for Plus, Pro, and Business subscribers and performed a one-time usage reset so everyone starts with extra headroom. That doesn’t mean GPT-5.6 Sol is suddenly infinite — it just means you won’t be yanked offline after that five-hour rolling block while the company works on capacity and efficiency improvements.

  • Plus, Pro, and Business plans had their five-hour restriction removed for now.
  • OpenAI reset current usage values, giving you more immediate allowance.
  • Local messages (like Codex work) and cloud tasks still share the same usage pool.
  • It’s a temporary measure — the old limits aren’t necessarily gone forever.

Why this matters (and what probably happened)

Demand spiked, so OpenAI chose two quick levers: clear the short-term cap and tune the model to be more efficient. The company said it’s rolling out changes to make GPT-5.6 Sol consume less usage per task, which likely means reduced token consumption or smarter internal batching. In practice, that gets you more time and throughput before you hit whichever limit is next in line.

Here are a few practical takeaways for people who actually use the model:

  • If you do lots of coding or run chains of agentic tasks, you probably feel noticeably less interrupted for now.
  • Watch your usage dashboard — the reset gives a buffer but won’t make the model free forever.
  • If you rely on local messages and cloud tasks, remember they draw from the same shared allowance, so heavy usage in one area affects the other.
  • Consider small efficiency wins on your side: shorter prompts, selective context, or throttling non-critical requests to stretch the allowance further.

In short: it’s a helpful pause button. OpenAI’s giving users a bit more breathing room while it tweaks GPT-5.6 Sol to be less greedy with tokens. Enjoy the smoother ride, but don’t assume the tap is permanently on — plan and monitor usage like the responsible, mildly paranoid pro you are.