August 6, 2026, OpenAI officially rolled out a major update to ChatGPT. Free users (including the Go plan) now default to the GPT‑5.6 Luna model, text conversations are no longer limited and can be used without restriction; Plus and Pro users receive an upgraded GPT‑5.6 Sol optimized for chat, with significantly improved factual accuracy and response quality. OpenAI CEO Sam Altman confirmed on X: "5.6 Sol performs much better in chat, and free users can also chat by text without limits!" The features are being rolled out to all users starting this week.
Update details
Free and Go users
- The default model switches from GPT‑5.5 to GPT‑5.6 Luna (the efficient, lightweight tier).
- Text conversations no longer have usage limits, enabling unlimited chat.
- A new "Think" button has been added so users can enable stronger reasoning for complex queries when needed.
- File uploads, images, voice, and other tools remain subject to existing limits to prevent abuse.
Internal evaluations show that GPT‑5.6 Luna reduces the factual error rate by about 62% compared with GPT‑5.5 Instant.
Plus and Pro users
- GPT‑5.6 Sol in ChatGPT receives targeted optimizations for everyday conversation, quick Q&A, web research, advice, planning, writing, and decision‑making scenarios.
- A new reasoning‑strength slider lets users adjust the model's depth of thought.
- Responses are more focused and better aligned with prompt needs, with reduced unnecessary formatting and redundancy.
- In internal tests, GPT‑5.6 Sol lowered the factual error rate under finance, medical, and legal prompts by about 68% compared with GPT‑5.5 Instant.
Note that this optimization applies only to the Chat experience in ChatGPT. The GPT‑5.6 Sol variants powering ChatGPT Work and Codex remain unchanged.
The update has begun rolling out on web, iOS, Android, and desktop; unlimited text for free users and the Think button are expected to be fully available next week.
Technical and product background
The GPT‑5.6 series (Sol, Terra, Luna) became broadly available in July 2026 and is positioned as flagship, balanced, and efficient tiers respectively. This update further strengthens the tiered strategy: Luna focuses on cost‑effectiveness and wide accessibility, while Sol receives targeted tuning for high‑frequency chat scenarios rather than simply applying general flagship capabilities.
OpenAI emphasized that the change reflects its ongoing effort to "get better models into more people's hands." ChatGPT's weekly active users have surpassed 1 billion, and upgrades to the free tier should help expand the user base further.
Possible impacts: user experience and market competition
1. Impact on free users Unlimited text conversations plus a stronger default model will significantly improve the experience for casual users and reduce "quota anxiety." The Think button also gives free users a better tool for handling complex problems. This may further solidify ChatGPT's consumer‑market lead.
2. Impact on paid users Sol's targeted optimizations make everyday chat more accurate and concise, reducing overthinking or misinformation. The slider increases controllability, letting users adapt to task complexity. For users who rely on ChatGPT for work and life tasks, the experience will be smoother.
3. Impact on OpenAI's business model Broadening capabilities in the free tier may put pressure on paid conversion, but higher engagement and word‑of‑mouth could attract more users into the ecosystem. The tiered model strategy also helps balance compute costs with user growth.
4. Impact on industry competition The move may prompt rivals such as Anthropic and Google to accelerate free‑ or low‑cost tier upgrades. The combination of unlimited text and high‑quality models will raise the industry's focus on accessibility. Open‑source and domestic models may still have cost advantages but will need to catch up on product polish and ecosystem integration.
5. Potential challenges
- Unlimited use could increase compute and content‑safety pressure and will require strict abuse prevention.
- Differences between chat‑only Sol and the Work/Codex variants may create inconsistency across user scenarios.
- Although factual accuracy has improved significantly, it is not perfect; users should still cross‑check information in high‑risk domains.
Industry insight: from "capability race" to "experience accessibility"
This update signals a shift from simply competing on flagship performance to finer‑grained scenario optimization and large‑scale accessibility. Free users gain near‑cutting‑edge default models and unlimited text, while paid users enjoy a more precise everyday experience—reflecting a product approach of "tiered service, broad improvement."
For general users, this is a tangible benefit; for developers and enterprises, it is a window into how OpenAI balances cost, safety, and growth. As features fully roll out, ChatGPT's everyday accessibility will further improve, moving AI assistants closer to being available to everyone.
▶️ Enjoyed this article? Creating content is hard work!
Follow our YouTube channel:
AI Rosetta
Your likes, subscriptions, and shares are our greatest support! 🙏
More AI news at: https://www.airosetta.com