The change in access
OpenAI is removing text-based rate limits for ChatGPT accounts on the free and Go plans. The shift begins next week. Users on those tiers will be able to continue text conversations without the previous caps that cut sessions short after a set number of exchanges.
File uploads, image handling, and other multimodal features keep their existing restrictions. The company is also introducing a Think button for these users that triggers higher reasoning steps on difficult prompts. The same button will appear across the free and Go interfaces.
Model updates across tiers
Free and Go accounts receive GPT-5.6 Luna as the new default model. It replaces GPT-5.5 Instant. MacRumors reported the switch directly from OpenAI’s announcement. The newer model is positioned as more capable for the everyday tasks these users perform.
Plus and Pro subscribers receive GPT-5.6 Sol instead. OpenAI states that Sol improves factual reliability and produces more focused replies. Both model lines include a new slider that lets users adjust how much reasoning effort the system applies to any given response.
Output formatting has been adjusted on the new models. The system now supplies only the information required by the query and drops unnecessary formatting that earlier versions sometimes added. This change applies to both the free-tier Luna model and the paid Sol model.
Limits that remain in place
Text messages alone lose their rate caps for free and Go users. Any message that includes a file upload or an image stays under the prior limits. The distinction shows OpenAI is separating compute costs between lightweight text turns and heavier multimodal requests.
The Think button gives free and Go users an on-demand path to deeper processing without changing the baseline limits on other features. TechCrunch noted the button alongside the removal of text rate limits. The Verge confirmed the same timing and scope.
Reactions and positioning
OpenAI has not released usage numbers or projected cost impact from the change. No third-party analysts have published independent estimates yet. The company frames the updates as refinements to model quality and output control rather than a broad expansion of service capacity.
Paid users see targeted improvements in factual grounding and response focus. Free and Go users gain longer text sessions and a newer default model. The split in treatment between text and other inputs remains consistent with earlier tier distinctions.
Why it matters
The move lowers friction for routine text use on free and Go accounts, which could increase daily engagement without raising the cost floor for heavier features. By keeping multimodal limits intact, OpenAI continues to meter the more expensive inference paths while testing wider text access. The model split between Luna and Sol also signals that OpenAI is willing to run parallel defaults rather than push every user onto the same weights.
For developers and power users already on paid plans, the Sol refinements and thought slider offer incremental control rather than new capabilities. Free users receive a clearer upgrade path on text alone. Over time the pattern suggests OpenAI will manage growth by carving out narrow expansions like unlimited text while protecting margins on everything else. The result is a more segmented product that rewards paid tiers with reliability tweaks and keeps free access bounded by input type.
---
Sources:
{"word_count": 682, "expanded_from": "previous_408_word_draft", "sources_used": 3}
No comments yet