OpenAI Cuts Free ChatGPT Chat Limits, Adds GPT-5.6

OpenAI is removing rate limits on text-based chats for Free and Go tier ChatGPT users, alongside the rollout of two new models — GPT-5.6 Luna and GPT-5.6 Sol — as the platform crosses 1 billion weekly users.

What's changing

According to the report, OpenAI is making several updates in quick succession:

Unlimited text chats: Free and Go users will no longer face rate limits on text-based conversations, though limits remain in place for file uploads, images, voice, and image generation across all tiers.
New default model: GPT-5.6 Luna is replacing GPT-5.5 as the default model for Free and Go users. Internally, OpenAI reports Luna produces 62% fewer factual errors than GPT-5.5-Instant.
Upgraded model for paid tiers: GPT-5.6 Sol is now available to Plus and Pro subscribers, with an internally reported 68% fewer factual errors than GPT-5.5-Instant. Sol is designed to be more reliable on dates, numbers, sources, rules, and assumptions, and delivers more direct responses with tighter formatting.
New reasoning controls: Free and Go users get a "Think" button to select higher reasoning power for complex questions. Plus and Pro users get a thinking slider to adjust how much reasoning the model applies to a given answer.

Timeline

Per the report's timeline:

This week: The default model for Free and Go users upgraded to GPT-5.6 Luna.
Thursday / today (sources differ): Updated GPT-5.6 Sol and the thinking slider became available to Plus and Pro users.
Next week: Unlimited text chats and the new Think button roll out to Free and Go users.

Sources differ on the exact timing of the Sol rollout — TechCrunch states it became available to Plus and Pro users "today," while The Verge states it and the slider would be available "starting Thursday." It's unclear whether these refer to the same date.

Scale context

The changes land as ChatGPT has crossed 1 billion weekly users, underscoring the scale at which OpenAI is now operating consumer-facing features.

Risks and open questions

The report flags several unresolved points:

Removing chat limits for free users could raise OpenAI's compute costs without a clear corresponding revenue increase.
Wider free access may increase server load, potentially affecting reliability or response times.
No details are available on why OpenAI is removing text limits now, what the previous limits were, or how the change affects infrastructure costs at 1 billion weekly users.
There's no comparison of Luna vs. Sol capabilities beyond the factual-error reduction figures, and no mention of pricing changes across Plus, Pro, or Go tiers alongside these updates.

Why founders should care

For early-stage founders building on or around ChatGPT, this update carries several plausible implications, though the report doesn't confirm exact outcomes:

Engagement potential: Unlimited free-tier text chats could likely increase user engagement and retention for products built on top of ChatGPT, as friction from rate limits disappears.
Reliability gains: The reported drops in factual error rates (62% for Luna, 68% for Sol versus GPT-5.5-Instant) suggest founders integrating ChatGPT outputs via API or web tools may see improved output reliability — though these figures are based on OpenAI's internal evaluations, not independent testing.
Cost/quality tuning: The new Think button and thinking slider may give founders more granular control over reasoning depth, which could be useful for tiering product features around cost and quality tradeoffs.
Competitive scale: With ChatGPT now past 1 billion weekly users, startups building consumer AI products should probably factor in the platform's growing reach when assessing competitive dynamics.

Founders relying on ChatGPT's free tier for prototyping or internal tools may want to monitor whether the removal of chat limits is sustained, given the open question of how OpenAI absorbs the associated compute costs at this scale.