OpenAI Expands GPT-3.5 Access With Unlimited Chats for All Users

Key Takeaways

• OpenAI has significantly expanded access to its advanced AI models, rolling out GPT-5.6 to all user tiers.
• Free and Go users now benefit from the GPT-5.6 Luna model and will soon receive unlimited text chat capabilities.
• Paid subscribers (Plus and Pro) gain access to the GPT-5.6 Sol model, which offers unified reasoning and user-controlled computational effort.
• Both new models demonstrate substantial improvements in factual accuracy, with reductions in errors by up to 68%.
• Enhanced safety measures, particularly for users under 18, have also been implemented.

OpenAI Democratizes Advanced AI with Unlimited Chats and Enhanced GPT-5.6 Models

OpenAI has announced a significant upgrade to its ChatGPT service, making its latest GPT-5.6 models available across all user tiers and, notably, removing text chat limitations for free users. This strategic move aligns with the company’s stated mission to ensure advanced artificial intelligence benefits a wider global audience.

Tiered Model Rollout

The update introduces two distinct models tailored to different user requirements: GPT-5.6 Sol, designed for Plus and Pro subscribers, and GPT-5.6 Luna, which will power the Free and Go tiers.

Improvements for Paid Subscribers

For ChatGPT Plus and Pro subscribers, the GPT-5.6 Sol model integrates instant and deep reasoning capabilities into a single, cohesive experience. This eliminates the previous distinction between quick responses and a more analytical “thinking mode,” which often led to an inconsistent conversational flow.

Paid users also gain granular control over the AI’s processing power. A new slider allows them to adjust the computational resources dedicated to each query, enabling optimization for tasks ranging from rapid everyday inquiries to complex coding, strategic planning, or in-depth research.

OpenAI says the model now delivers more concise and focused answers, reduces superfluous formatting, and proactively corrects user misconceptions rather than simply agreeing with incorrect premises.

Major Gains for Free and Go Users

The most substantial benefits of this rollout are directed at Free and Go tier users. The GPT-5.6 Luna model is being deployed as the default for these tiers this week. Crucially, starting next week, these users will gain access to unlimited text chats, removing previous rate limits that restricted usage.

Additionally, free users will receive a new “Think” button, enabling them to prompt the AI for more intensive reasoning on challenging questions. While this feature is subject to anti-abuse protocols, it significantly enhances the capabilities available without a subscription. However, certain limitations, such as those on file uploads and image generation, will remain in place for free users.

Enhanced Accuracy and Safety Measures

A headline improvement across both new models is a significant boost in factual accuracy. Internal evaluations across diverse domains, including financial, medical, and legal prompts, revealed a substantial reduction in factual errors. GPT-5.6 Luna demonstrated approximately a 62% decrease in errors, while GPT-5.6 Sol achieved an impressive 68% reduction compared to the preceding GPT-5.5 Instant model. This addresses a persistent criticism of large language models concerning “hallucinated” facts, dates, and figures.

OpenAI has also detailed expanded safety protocols in its accompanying system card. These measures specifically target users believed to be under 18, implementing restrictions on romantic roleplay, age-restricted content, and prompts that might encourage inappropriate real-world support connections.

It is important to note that the chat-optimized version of GPT-5.6 Sol will not be extended to OpenAI’s Work or Codex products, which will maintain their current model configurations. This strategic differentiation underscores OpenAI’s ongoing commitment to democratizing advanced AI capabilities while simultaneously enhancing factual reliability across its consumer-facing product suite.

Disclaimer: HackersRadar reports on cybersecurity threats and incidents for informational and awareness purposes only. We do not engage in hacking activities, data exfiltration, or the hosting or distribution of stolen or leaked information. All content is based on publicly available sources.