--- title: OpenAI upgrades ChatGPT models across all tiers, brings... description: OpenAI is rolling out GPT-5.6 model upgrades across all ChatGPT tiers. Free and Go users get GPT-5.6 Luna as the new default, with unlimited text chats, no... canonical: https://daily.dev/posts/openai-upgrades-chatgpt-models-across-all-tiers-brings-unlimited-text-chats-to-free-users-3ybfz5wai twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: OpenAI upgrades ChatGPT models across all tiers, brings unlimited text chats to free users | daily.dev og:description: OpenAI is rolling out GPT-5.6 model upgrades across all ChatGPT tiers. Free and Go users get GPT-5.6 Luna as the new default, with unlimited text chats, no... og:url: https://daily.dev/posts/openai-upgrades-chatgpt-models-across-all-tiers-brings-unlimited-text-chats-to-free-users-3ybfz5wai og:image: https://api.daily.dev/og/posts/3ybfZ5WaI.png og:image:alt: OpenAI upgrades ChatGPT models across all tiers, brings unlimited text chats to free users og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # OpenAI upgrades ChatGPT models across all tiers, brings unlimited text chats to free users **[Collections](https://daily.dev/sources/collections)** · 4 min read · 2 upvotes · 0 comments ## Summary OpenAI is rolling out GPT-5.6 model upgrades across all ChatGPT tiers. Free and Go users get GPT-5.6 Luna as the new default, with unlimited text chats, no message caps, and a new 'Think' button for deeper reasoning — Luna reportedly cuts factual errors by 62% vs GPT-5.5 Instant. Plus and Pro users receive GPT-5.6 Sol, which unifies instant and deep reasoning into a single model with a thinking effort slider, reducing factual errors by 68% in internal finance, medicine, and law evaluations. The Sol update eliminates the jarring model-switch behavior previously experienced when changing reasoning effort levels. ## Content OpenAI has reshuffled how ChatGPT picks its brain, depending on who's paying (or not paying) for it. The headline change: Plus and Pro subscribers now get a single model, GPT-5.6 Sol, for both quick answers and deep reasoning. That's a bigger deal than it sounds. Previously, switching from Instant to a higher reasoning effort could feel like talking to a different assistant, since the tone and behavior shifted along with the model swap. Now there's one model underneath, with a new slider that controls how much thinking effort ChatGPT puts in, from Instant up to High. Sam Altman put it simply: "5.6 Sol much better in chat now." Greg Brockman called the unification "a very significant step towards making ChatGPT simpler and more intuitive." Free and Go users aren't getting Sol, but they're getting something arguably more useful day-to-day: unlimited text chats, powered by a new model called GPT-5.6 Luna, plus a "Think" button for harder questions. Daily limits are gone (subject to the usual abuse protections OpenAI always mentions in the fine print). Given that ChatGPT passed a billion monthly users back in May, removing friction for free users looks like a fairly direct response to competitive pressure from Claude and Gemini. On accuracy, OpenAI's internal evaluations claim GPT-5.6 Sol produces 68% fewer responses with at least one factual error compared to GPT-5.5 Instant, tested across finance, medicine, and law. Luna comes in at a 62% reduction over the same baseline. Worth flagging: these are OpenAI's own numbers, compared against the previous Instant model rather than the previous Sol version, and the methodology isn't fully reproducible from what's been published. Take the percentages as directional rather than gospel. A few things stay unchanged. Sol in Codex and in ChatGPT Work is not affected by this update, so if you're testing prompts in consumer ChatGPT before deploying them elsewhere, you may actually be testing a different model than what ends up in production. OpenAI is also rolling in stricter content protections for users it believes are under 18, though loosening usage caps for everyone else does raise the obvious question of how those two things square up. Separately, OpenAI previewed an "Ultrafast" mode for Sol, built with Cerebras, that reportedly runs up to 14x faster. It's currently limited to a small group of API customers, with others able to sign up for updates. OpenAI says the goal right now is learning where that speed actually matters before deciding how to build it into products more broadly. **How it's landing with developers** Reactions from people actually using Sol have been a mix of genuine excitement and pointed skepticism. One researcher, scaling01, compared reasoning traces between Sol and Anthropic's Opus 4.8 and came away with a striking read: Opus seems to operate at a higher level of abstraction, while Sol grinds through the problem directly, throwing out hypotheses until something sticks. Their take was that Sol's approach is more efficient and arguably more scientific in method, and that OpenAI's trajectory this year has them more confident the company could retake the lead in frontier models soon. The same person flagged a real risk, though: models trained this way are more likely to develop reasoning that's hard for humans to interpret, which makes monitoring them harder. On the practical side, developers are reporting good results on messy real-world tasks. One person used Sol to audit a 3-million-entry US database and found it outperformed Claude Opus 5 as a reviewer, though it's worth noting that having one model grade another's output isn't the same as an independent ground-truth benchmark. A Columbia PhD candidate says they solved six open Erdős math problems in five days using Sol's ## Questions this post answers ### What is GPT-5.6 Luna and how does it differ from GPT-5.5 for free ChatGPT users? GPT-5.6 Luna replaces GPT-5.5 as the default model for ChatGPT Free and Go users, bringing unlimited text chats with no message caps and a new 'Think' button for harder questions requiring more reasoning. Internal evaluations show Luna has a 62% lower factual error rate than GPT-5.5 Instant. _Developers tracking ChatGPT model changes and their impact on AI-assisted workflows follow updates like these on daily.dev._ ### What does the GPT-5.6 Sol thinking effort slider do for Plus and Pro users? The thinking effort slider in GPT-5.6 Sol lets Plus and Pro users control how much reasoning the model applies to a response, all within a single unified model. Previously, switching between instant and deep reasoning modes felt like switching to a different model entirely, with noticeable tone and behavior shifts. Sol eliminates that discontinuity while reducing factual errors by 68% vs GPT-5.5 Instant in finance, medicine, and law evaluations. _Teams deciding how to integrate ChatGPT into professional workflows track model capability changes like this on daily.dev._ ## Community take How the wider developer community reacted, aggregated from 2 discussions and 111 comments across hackernews, x (as of 2026-08-13). **TL;DR:** There's essentially no usable discussion content provided for this release, so no real community sentiment can be determined. Improving GPT-5.6 Sol in ChatGPT—and expanding access for free users: The community is cautiously positive about free users getting access to Luna and the 'Think' toggle, but there's significant frustration over the proliferation of model tiers and reasoning knobs, confusion about the model naming/versioning, and debate over whether Luna is actually an upgrade or downgrade for free users. **Sentiment:** 30% positive · 45% mixed · 25% skeptical **The case for** - Giving free users access to reasoning ('Think' toggle) is seen as a meaningful improvement with broad societal impact. - Unlimited Luna for free users is viewed as a genuinely generous move, especially given Luna's improved quality over prior 'Instant' models. - Improvements to Sol's accuracy and consistency are welcomed by paying users who rely on it daily. **The pushback** - The explosion of model tiers (Luna, Terra, Sol, Pro, Ultra, Max, Xhigh, etc.) is seen as overwhelming and poorly explained, leaving users unsure which combination to use. - Some argue this is actually a downgrade for free users, since Luna is nano-equivalent and worse than the GPT-5.5 mini they previously received. - Auto-routing was tried before and failed, and the reasoning slider is seen as a poor substitute for the model just asking clarifying questions. - Recent ChatGPT UX bugs (auto-formatting, copy-paste issues) undermine the appeal of the free tier improvements. - GPT models are criticized for being RLHF-tuned toward engagement over directness, giving verbose non-answers instead of concise replies. **By community** - hackernews (mixed): Broadly receptive to the free-tier expansion but deeply frustrated by model proliferation, naming confusion, and the reasoning-level UX, with a lively side debate on whether Luna is actually a step up or down for free users. - x (mixed): No substantive replies were provided, so no clear lean can be drawn. **Hottest debate:** Improving GPT-5.6 Sol in ChatGPT—and expanding access for free users: Whether the move to unlimited Luna is a genuine upgrade or a downgrade for free users, given that Luna is described as nano-equivalent and worse than the GPT-5.5 mini previously offered. **Open questions** - Is there actually a GPT-5.6 Instant model, or is Luna simply replacing the Instant tier entirely? - How does OpenAI plan to monetize the free tier at scale given LLM inference costs? - Will auto-routing ever work reliably, or does detecting query complexity require a complex model in itself? - What is the actual performance difference between Terra, Luna, and Sol for real-world coding tasks? **Highlights** > Auto-effort and similarly auto model routing suffer from a halting problem sort of issue: you don’t reliably know if a request is complex unless you use a complex model to make the decision. > — [nojs on hackernews](https://news.ycombinator.com/item?id=49203662) > Giving free ChatGPT users access to reasoning (the 'Think' toggle) will have a broader impact on the world than every new paid model and coding agent combined. > — [heaney-555 on hackernews · 4 comments](https://news.ycombinator.com/item?id=49201557) > I have the opposite problem. I'm not well-calibrated on when I'd want lower reasoning than what's available to me (and how to compare that to lower-tier models). OpenAI now has Luna, Terra and Sol, each at Low, Medium, High and Xhigh, with Pro/Ultra depending on harness and plan. That's ~15 possible combinations of model and reasoning level, and there isn't a satisfactory explanation of which one you want for any particular task. I feel that work is basically split into two tiers, hard (which requires a good model and lots of reasoning by definition) and relatively easy (which won't consume much of my limits despite a great model and reasoning, so I may just as well keep it on Sol High). > — [miki123211 on hackernews · 2 comments](https://news.ycombinator.com/item?id=49204197) > They must really be feeling the commoditization pressure. I'm not sure what the way out of this is, ChatGPT and Claude are still good products, but they are not necessarily premium products anymore. I expect a few things to happen in the next year: 1) Exclusive MCP server deals/API integrations 2) Significant switch to B2B marketing, even moreso than we've seen before, with API interfaces being paid and chat-client interfaces becoming more and more free, perhaps just with limits more on integrations or data visualization/analysis 3) US restrictions on B2B contracts with non-US hosted models that do any sort of contracting with the government Obviously there's a bunch of stuff I'm not foreseeing. But it really does feel like the bottom of the market is collapsing into free. I assume OpenAI and Anthropic think their next generation of models will restore their halo tier status and that the cash burn is justified to just get there, but this has to really mess up IPO plans. > — [colingauvin on hackernews · 1 comments](https://news.ycombinator.com/item?id=49201031) > This is actually a downgrade for free users since currently it uses GPT-5.5 for a few messages before it drops you down to GPT-5.5-mini. Now it always uses a model worse than Mini (Luna is nano-equivalent, "It roughly corresponds to the nano model tier used in earlier GPT-5 families." https://developers.openai.com/api/docs/models/gpt-5.6-luna ). I guess it's a bit better with Thinking though. They should use Terra for a few messages first before dropping down to Luna. And image inputs are still limited. > — [sunaookami on hackernews](https://news.ycombinator.com/item?id=49202667) **Source threads** - [hackernews](https://news.ycombinator.com/item?id=49199357) · 143 points · 111 comments - [x](https://x.com/testingcatalog/status/2087951653374677152) · 0 points · 0 comments - [x](https://x.com/jxnlco/status/2087956973929324835) · 0 points · 0 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#openai](https://daily.dev/tags/openai), [#chatgpt](https://daily.dev/tags/chatgpt) [View this post on daily.dev](https://daily.dev/posts/openai-upgrades-chatgpt-models-across-all-tiers-brings-unlimited-text-chats-to-free-users-3ybfz5wai) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"OpenAI upgrades ChatGPT models across all tiers, brings unlimited text chats to free users","url":"https://daily.dev/posts/openai-upgrades-chatgpt-models-across-all-tiers-brings-unlimited-text-chats-to-free-users-3ybfz5wai","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/openai-upgrades-chatgpt-models-across-all-tiers-brings-unlimited-text-chats-to-free-users-3ybfz5wai"},"datePublished":"2026-08-06T19:22:50.803Z","dateModified":"2026-08-13T17:39:28.657Z","description":"OpenAI is rolling out GPT-5.6 model upgrades across all ChatGPT tiers. Free and Go users get GPT-5.6 Luna as the new default, with unlimited text chats, no...","isAccessibleForFree":true,"articleSection":"Collections","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Collections","logo":"https://media.daily.dev/image/upload/s--fk_6ycEi--/f_auto,q_auto/v1780996001/logos/collections?_a=BAMAMiWQ0","url":"https://daily.dev/sources/collections"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/openai-upgrades-chatgpt-models-across-all-tiers-brings-unlimited-text-chats-to-free-users-3ybfz5wai","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":2},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,openai,chatgpt","timeRequired":"PT4M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Collections","item":"https://daily.dev/sources/collections"},{"@type":"ListItem","position":3,"name":"OpenAI upgrades ChatGPT models across all tiers, brings unlimited text chats to free users"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/openai-upgrades-chatgpt-models-across-all-tiers-brings-unlimited-text-chats-to-free-users-3ybfz5wai#faq","mainEntity":[{"@type":"Question","name":"What is GPT-5.6 Luna and how does it differ from GPT-5.5 for free ChatGPT users?","acceptedAnswer":{"@type":"Answer","text":"GPT-5.6 Luna replaces GPT-5.5 as the default model for ChatGPT Free and Go users, bringing unlimited text chats with no message caps and a new 'Think' button for harder questions requiring more reasoning. Internal evaluations show Luna has a 62% lower factual error rate than GPT-5.5 Instant. Developers tracking ChatGPT model changes and their impact on AI-assisted workflows follow updates like these on daily.dev."}},{"@type":"Question","name":"What does the GPT-5.6 Sol thinking effort slider do for Plus and Pro users?","acceptedAnswer":{"@type":"Answer","text":"The thinking effort slider in GPT-5.6 Sol lets Plus and Pro users control how much reasoning the model applies to a response, all within a single unified model. Previously, switching between instant and deep reasoning modes felt like switching to a different model entirely, with noticeable tone and behavior shifts. Sol eliminates that discontinuity while reducing factual errors by 68% vs GPT-5.5 Instant in finance, medicine, and law evaluations. Teams deciding how to integrate ChatGPT into professional workflows track model capability changes like this on daily.dev."}}]} ```