OpenAI is launching its powerful GPT-5.6 model series to the public on Thursday following approval from the US government. The release comes after weeks of restricted testing with US-based partners and marks a significant moment in the AI race between OpenAI and rival Anthropic. The new models represent a shift in how the government is managing access to cutting-edge AI technology, balancing national security concerns with the need to keep advanced tools available to the public.

What Are the Three New GPT-5.6 Models?

OpenAI is rolling out GPT-5.6 in three distinct tiers, each designed for different use cases and budgets. The company structured the release to give users options rather than forcing everyone into a one-size-fits-all model. This tiered approach reflects a broader industry trend of offering capability levels that match real-world needs.

• Sol: OpenAI's new flagship model, designed for the most demanding tasks requiring maximum capability and performance
• Terra: A mid-range version for everyday work, priced at half the cost of its predecessor GPT-5.5, making it accessible to more users and businesses
• Luna: A fast, low-cost option optimized for speed and affordability, ideal for simple queries and routine tasks

The pricing strategy for Terra is particularly noteworthy. By cutting the cost in half compared to GPT-5.5, OpenAI is signaling aggressive competition with Anthropic, which recently restored access to its own powerful models, Fable 5 and Mythos 5, after similar government restrictions were lifted.

Why Did the US Government Restrict These Models in the First Place?

Both GPT-5.6 and Anthropic's Mythos series drew national security concerns because of their unprecedented ability to identify software vulnerabilities, weaknesses in code that hackers could exploit. The Trump administration initially wanted to carefully control where and how these models could be released, fearing that unrestricted access could pose cybersecurity risks.

However, the administration's approach shifted. Rather than imposing long-term restrictions, officials decided to develop a repeatable framework for evaluating future models. OpenAI stated in June that it was working with Washington "to develop the cyber Executive Order framework and a repeatable process for future model releases." This suggests the government is moving toward a more systematic approach rather than ad-hoc restrictions on each new model.

How to Get the Most Out of Your AI Interactions

While OpenAI's new models are more capable, the quality of your results still depends heavily on how you interact with them. One journalist who tested hundreds of AI prompts discovered a simple technique that dramatically improved ChatGPT's responses across all models.

• Ask for Clarification First: Add a sentence to your prompts asking the AI to request clarifying questions before answering. Instead of guessing what you mean, the model pauses to understand your actual intent, reducing the need for follow-up corrections
• Provide Context for Complex Tasks: When working on important projects like vacation planning, email writing, or career advice, give the AI permission to ask questions about your specific situation, budget, preferences, and constraints rather than forcing it to make assumptions
• Adjust for Simple Queries: For straightforward questions like "What is the capital of Portugal?", modify your approach to ask the AI to only request clarifying questions if necessary, preventing unnecessary delays while still catching important gaps

The technique works across ChatGPT, Claude, Gemini, and most other AI chatbots. One effective prompt addition is: "Before answering, ask up to three clarifying questions only if they're necessary to avoid making incorrect assumptions. Otherwise, answer immediately and mention any assumptions you made." This small change shifts the conversation from "guess what I mean" to "help me figure out the best answer".

What Does This Mean for the AI Competition?

The stakes are enormous. Both OpenAI and Anthropic have filed confidential IPO documents with US regulators and are targeting public listings at valuations approaching one trillion dollars. The commercial pressure to lock in customers is intense, which explains why OpenAI is aggressively pricing Terra at half the cost of its predecessor.

Anthropic's recent restoration of access to Fable 5 and Mythos 5 set the stage for OpenAI's broader launch. The two companies are now in direct competition not just on capability, but on availability and pricing. OpenAI's decision to release three tiers simultaneously suggests the company is trying to capture market share across different customer segments, from enterprises willing to pay for Sol's maximum performance to budget-conscious users who prefer Luna's speed and affordability.

The government's shift toward a repeatable approval framework rather than permanent restrictions also signals confidence that both companies can be trusted to release powerful models responsibly. This could accelerate the pace of AI releases in the coming months, as companies no longer face indefinite delays waiting for case-by-case government review.