TechCrunch: OpenAI removes chat limits for free ChatGPT users

On August 6, 2026, OpenAI announced that free and Go-tier ChatGPT users would receive unlimited text-based chats — removing the message caps that had previously defined the boundary between free and paid tiers. The rollout was set to begin the following week. Along with removing the limit, OpenAI upgraded the default model for free users to GPT-5.6 Luna, a version reported to produce 62% fewer factual errors than its predecessor. Paid subscribers on Plus and Pro plans receive access to GPT-5.6 Sol, a higher-reasoning variant.

The structural change in how OpenAI is thinking about freemium is notable. With 1 billion weekly users, the company is moving away from constraining access through usage quotas and toward differentiating through capability depth — reasoning sliders, file and image processing, voice, and advanced thinking modes remain gated behind paid plans.

For product managers building on or competing with AI platforms, the announcement carries a few implications. Usage limits have historically been a mechanism for converting free users to paid, and their removal signals that OpenAI believes engagement and habit formation at the free tier will produce more conversion than friction from hitting limits. This is a bet on user retention over artificial scarcity, and it changes the calculus for any product that has built its own freemium model around similar constraints.

More broadly, upgrading the free tier to a capable model while keeping a distinct premium tier for reasoning-intensive work suggests a segmentation strategy: casual users stay on the free product, while professionals and enterprises pay for the capability they actually need. It is a cleaner split than quota-based differentiation, and one that other AI platforms may follow.