OpenAI has released its latest family of models: GPT-5.6 Luna, Terra, and Sol. They come in three sizes — from the budget-friendly Luna to the powerhouse Sol — and they represent a significant shift in how AI companies are thinking about pricing and capability. For Irish business owners trying to make sense of AI costs, this is worth understanding.
The headline numbers are straightforward. Luna costs $1 per million input tokens and $6 for output. Terra is $2.50 and $15. Sol costs $5 and $30. For comparison, Anthropic’s Claude Opus runs at $5 and $25, and their top-end Fable 5 is $10 and $50. But raw per-token pricing only tells part of the story, because these models can use dramatically different amounts of reasoning tokens to solve the same problem.
Three Models for Three Different Needs
OpenAI has structured the GPT-5.6 family so businesses can pick the model that matches the complexity of the task. If you just need a quick answer or a straightforward document summary, Luna will do the job at a fraction of the cost. If you’re tackling complex analysis or building AI agents that need to reason through multi-step problems, Sol is the better choice but costs more per query.
This tiered approach is a response to what businesses have been telling AI companies: one model does not fit all. Many small businesses have been paying for top-tier AI capability when they only needed basic assistance. The GPT-5.6 family gives them a cheaper option that’s still very capable.
All three models share some impressive specs: a knowledge cutoff of February 2026, a million-token context window, and the ability to output up to 128,000 tokens in a single response. The million-token context window is particularly interesting for businesses — it means you could feed an entire year’s worth of documents into the model in one go.
New Features That Matter for Business
The GPT-5.6 release also comes with new capabilities baked into the API. Programmatic tool calling lets the model compose and run code as part of its reasoning process. Multi-agent support allows the AI to spin up sub-agents for parallel work – useful for tackling several business problems at once. And prompt cache breakpoints give businesses more control over costs by being explicit about where cached data starts and ends.
What’s most interesting from a business perspective is that OpenAI is claiming efficiency gains: GPT-5.6 Sol beats Claude Fable 5 on the Agents’ Last Exam benchmark by 13 points, and the smaller models achieve this at roughly one-sixteenth the cost. Whether or not benchmarks translate to real-world performance, the trend is clear — AI is getting both cheaper and more capable at the same time.
How to Choose the Right Model for Your Business
If you’re currently paying for a top-tier AI subscription, ask yourself what you’re actually using it for. If it’s mostly drafting emails, summarising documents, and basic research, Luna will likely handle those tasks at a fraction of the cost. If you’re building automated workflows, analysing complex data, or developing software, Terra or Sol might be worth the premium.
The key insight is that AI pricing is finally becoming flexible enough to match real business needs. You no longer have to buy the most expensive option to get good results. Luna at $1 per million input tokens is genuinely useful for everyday business tasks. That changes the calculation for any Irish small business owner wondering whether AI is worth the cost.
What to Watch Next
The competition between AI providers is intensifying. OpenAI’s tiered pricing follows similar moves from Anthropic and Google. For businesses, this is entirely good news: more options, lower prices, and better capability. If you’ve been waiting for the right moment to start using AI seriously in your business, the pricing is now favourable enough to begin experimenting without breaking the budget.