OpenAI Removes Text Chat Limits for Free ChatGPT Users, Launches GPT-5.6 Luna
OpenAI is removing text chat limits for Free and Go ChatGPT users, powered by a new GPT-5.6 Luna model with a 'Think' button for harder questions. The company also upgraded GPT-5.6 Sol for Plus and Pro users, claiming a 68% reduction in factual errors versus GPT-5.5-Instant.
OpenAI is removing usage limits on text-based chats for all ChatGPT users, including those on the free tier, the company announced Thursday. The change arrives as ChatGPT recently crossed 1 billion weekly users.
A new model called GPT-5.6 Luna will power the free experience, replacing GPT-5.5 as the default model for Free and Go tier users. Alongside the model swap, Free and Go users will get a new "Think" button that lets them request higher reasoning effort for complex questions. OpenAI said usage caps will remain in place for files, images, voice, and image generation — only unrestricted text chat is changing.
The rollout is staggered. Free and Go users will see the GPT-5.6 Luna default and the Think button appear this week. Unlimited text chats specifically go live next week.
Changes for Plus and Pro users
Paying subscribers are also getting updates. Plus and Pro users gain access to an upgraded GPT-5.6 Sol model, which OpenAI says is tuned for quick tasks — questions, web research, advice, planning, writing, and decision-making — and produces more compact answers. This version is already live today.
OpenAI clarified that this consumer-facing GPT-5.6 Sol is a distinct build from the GPT-5.6 Sol variant used in Codex and Work products, which remains unchanged.
Plus and Pro users also get a "thinking slider" that lets them manually adjust how much computational effort the model applies to a given query, tuning it up for multi-step problems or down for simple ones.
Accuracy claims
According to OpenAI's internal evaluations, factual errors dropped 62% for GPT-5.6 Luna and 68% for GPT-5.6 Sol compared to GPT-5.5-Instant. OpenAI has not published the evaluation methodology, dataset, or sample size behind these figures, so the numbers should be treated as company-reported claims rather than independently verified benchmarks.
No pricing changes were announced. ChatGPT Free remains free, and Plus/Pro/Go pricing tiers are unaffected by this update; pricing for API access to GPT-5.6 Luna or Sol has not yet been disclosed.
What this means
Removing text chat caps for free users is a meaningful shift in ChatGPT's business model — historically, rate limits have been one of the primary levers pushing free users toward paid tiers. With ChatGPT at 1 billion weekly users, OpenAI appears to be betting that broader unrestricted access drives more overall engagement and ad/data value (or future monetization paths) rather than direct subscription conversion. The move also puts pressure on competitors like Google's Gemini and Anthropic's Claude, which still meter free-tier usage more aggressively.
The claimed error-rate improvements — 62% and 68% — are significant if accurate, but without published benchmark methodology, they function more as marketing signal than verifiable performance data. The bifurcation of model naming (consumer Sol vs. Codex/Work Sol) also signals OpenAI is now maintaining multiple specialized variants under the same version label, which could create confusion for developers building on the API rather than the ChatGPT app.
Related Articles
Simon Willison's Pelican Benchmark Shows GPT-6 Astra Outperforming GPT-5.6 Sol at Every Reasoning Level
Developer Simon Willison ran his signature 'pelican riding a bicycle' SVG test on newly-accessed GPT-6 Astra across five reasoning levels, comparing results against GPT-5.6 Sol, Terra, and Luna. Even Astra's lowest reasoning setting reportedly beat every Sol output, though Astra costs roughly twice as much per token.
Pentagon Adds OpenAI's ChatGPT Mil and xAI's Grok for Government to GenAI.mil
The Pentagon has added OpenAI's ChatGPT Mil and xAI's Grok for Government to its GenAI.mil platform, which previously offered only Google Gemini. Anthropic's Claude remains excluded after a supply-chain risk dispute with the Trump administration.
OpenAI Lists GPT-6 Astra Pro on OpenRouter: Same Model, Higher-Compute Reasoning Mode
GPT-6 Astra Pro, now listed on OpenRouter, is the existing GPT-6 Astra model configured to run with reasoning.mode set to 'pro' for higher-quality output on complex tasks. It carries a 1M-token context window and tiered pricing from $5/$25 to $20/$100 per million input/output tokens depending on the serving tier.
OpenAI's GPT-6 Astra Cuts Hallucinations, But Indirect Prompt Injection Attacks Still Succeed 8.5% of the Time
OpenAI's new GPT-6 Astra model shows major improvements in hallucination rates and jailbreak resistance over predecessor GPT-5.6 Sol, according to OpenAI's system card. However, indirect prompt injection attacks hidden in documents still succeed 8.5% of the time in external testing by Gray Swan, down from 27% but still above rival Claude Opus 5's 4.8% rate.
Comments
Loading...