AI Pulse · Frontier Labs Watch

OpenAI just dropped message caps for free users. If you’re on the free or Go tier, you can now send unlimited text chats using GPT-5.6 Luna, the new default model. Limits still apply to file uploads, images, and tools, but the core conversation engine is open. Starting the week of August 10, everyone gets a “Think” button that triggers extended reasoning when you need the model to work through something harder. Paid subscribers on Plus or Pro get access to an upgraded GPT-5.6 Sol with an effort slider that lets you dial up or down how much compute the model throws at a task.

This is a land grab. OpenAI is betting that unlimited free access to a capable model will lock in daily users before Google and Meta push their own consumer plays harder. They’re trading near-term free-tier revenue for distribution and habit formation. If you run a business that competes on customer engagement or needs to understand how AI adoption shifts user expectations, this matters. Your customers are about to have a free, always-on assistant in their pocket that handles a lot of what used to require a human or a paid tool.

For internal ops, the effort slider on Sol is the more interesting piece. Being able to tune how much reasoning a model applies to a given task means you can match cost to complexity instead of overpaying for simple queries or underpowering hard ones. That kind of control is exactly what we build into systems like the Omni Command Centre, where you route different tasks to different models and effort levels based on what the job actually needs.

You can read more in the official announcement. The short version is that free AI just got a lot more capable, and that changes the baseline for what people expect software to do.