While Anthropic spent this week mostly tinkering with Claude Code, OpenAI made a move toward the masses: the daily limit on free ChatGPT chats is gone. For anyone using ChatGPT for free, that’s the most tangible change in a long time.
Type without limits, on a better default model
Starting this week, Free and Go users get unlimited text chats — no more text rate limits. The important caveat: this applies to plain text. For file uploads, images, and other ChatGPT tools, the limits stay in place.
At the same time, GPT-5.6 Luna becomes the new default model for Free and Go. OpenAI puts a number on the improvement: Luna produces 62 percent fewer factual errors than the old GPT-5.5 Instant. So if you were cautious with the free version before, you now get not just more volume but more reliability too.
A button for thinking
The second piece arrives the following week: a Think button for free users. For trickier questions, it lets you request higher reasoning power — gated by abuse guardrails so nobody blows the reasoning budget. Reasoning on demand, instead of hiding it behind an expensive subscription.
For paying users, OpenAI sharpens GPT-5.6 Sol: more reliable with facts, more focused answers. And a new slider lets you decide how much thought ChatGPT puts into each response.
My take
This is less a model story than a strategic one. OpenAI is giving away exactly what most competitors keep behind the paywall: unlimited use of a good model and dial-able reasoning. The fight is shifting from “who has the best model?” to “who weaves themselves into the most people’s daily routine?”
For those of us working with Claude, it still matters. Pressure from the free side bleeds over — into prices, into limits, into the question of what a subscription even justifies anymore. Competition is annoying for the vendors, but pretty great for us.
Sources: OpenAI: Improving GPT-5.6 Sol in ChatGPT — and expanding access to GPT-5.6 Luna for free users, TechCrunch: ChatGPT brings unlimited text chats to free users