OpenAI is changing its usage limits in December: ChatGPT and Codex will share a single usage pool.
Here’s how it works: Starting in December, Plus, Pro, and Business subscribers will no longer have separate usage limits for ChatGPT and Codex. Instead, usage will be deducted from one shared pool. Usage is measured in tokens—the more advanced the model and the more complex the task, the more tokens it uses. The Work feature lets you delegate tasks like building websites, creating presentations, and operating web pages directly in ChatGPT, and you can see your remaining usage all in one place on the usage page in Settings. The good news is that unlimited Instant usage won’t be affected.
This change will have the biggest impact on two groups: First, people who regularly use both chat and coding heavily will now be drawing from the same pool, so they’ll need to budget their usage across both. Second, people who rely on Codex for major projects—the combination of advanced models and complex tasks will use up the pool fastest, and they could easily run out before the end of the month.
My take: OpenAI is rolling back the scope of “unlimited” usage. Codex’s previously more independent usage limit is now being folded into the shared pool, effectively putting a ceiling on heavy users. I suspect Plus subscribers will feel the change first after December, while Pro subscribers can hold out a bit longer thanks to their higher limits. For those of us who use ChatGPT for both chatting and coding, managing usage will become essential from December onward. It’s probably best not to pile casual chats and major projects onto the same day.
#AI Frontier
$BTC
Here’s how it works: Starting in December, Plus, Pro, and Business subscribers will no longer have separate usage limits for ChatGPT and Codex. Instead, usage will be deducted from one shared pool. Usage is measured in tokens—the more advanced the model and the more complex the task, the more tokens it uses. The Work feature lets you delegate tasks like building websites, creating presentations, and operating web pages directly in ChatGPT, and you can see your remaining usage all in one place on the usage page in Settings. The good news is that unlimited Instant usage won’t be affected.
This change will have the biggest impact on two groups: First, people who regularly use both chat and coding heavily will now be drawing from the same pool, so they’ll need to budget their usage across both. Second, people who rely on Codex for major projects—the combination of advanced models and complex tasks will use up the pool fastest, and they could easily run out before the end of the month.
My take: OpenAI is rolling back the scope of “unlimited” usage. Codex’s previously more independent usage limit is now being folded into the shared pool, effectively putting a ceiling on heavy users. I suspect Plus subscribers will feel the change first after December, while Pro subscribers can hold out a bit longer thanks to their higher limits. For those of us who use ChatGPT for both chatting and coding, managing usage will become essential from December onward. It’s probably best not to pile casual chats and major projects onto the same day.
#AI Frontier
$BTC
