On June 26, 2026, OpenAI announced its GPT-5.6 model family: three distinct AI models built for different workloads and price points. They are named Sol, Terra, and Luna, and together they represent the most significant structural shift in how OpenAI is packaging its AI to date.
Instead of one flagship with a smaller sibling, you now get a three-tier lineup where each model serves a specific job. That matters for business owners and marketers because the days of one-size-fits-all AI pricing are ending. The question is no longer “should we use AI” but “which model matches which task.”
Here is a straight breakdown of each one, and what it means for your operations.
GPT-5.6 Sol: The Flagship, Built for Demanding Work
Sol is the top of the stack. OpenAI describes it as its strongest model to date, and the benchmarks back that up. It sets a new record on Terminal-Bench 2.1, which tests command-line workflows requiring multi-step planning and tool coordination. It outperforms competing frontier models on that evaluation.
What makes Sol different technically is a new “max” reasoning mode that gives the model extended time to work through complex problems before responding, and an “ultra” mode that coordinates multiple subagents to tackle work that a single agent cannot handle efficiently. That is a meaningful jump from how AI tools have worked up to this point. Most current tools run one agent doing one thing. Sol can spin up parallel agents, divide a complex task, and synthesize the results.
For business owners, the most practical applications are in high-stakes document work, multi-step research, advanced coding, and competitive analysis where accuracy genuinely matters. If you are running a marketing team that currently uses AI for isolated tasks (write this, summarize that), Sol is built to chain those tasks together with less supervision.
Pricing: $5 per million input tokens / $30 per million output tokens via API.
Availability: Sol is currently in limited preview, accessible only to a small group of government-approved partner organizations. OpenAI plans to make it broadly available through ChatGPT, Codex, and the API in the coming weeks, though no specific public date has been set.
GPT-5.6 Terra: Everyday Performance at Half the Cost
Terra is the model most business owners and marketing teams will actually use day to day. OpenAI positions it as a balanced option for everyday work, with performance competitive with GPT-5.5, its previous frontier model, at roughly half the cost.
That pricing structure matters. GPT-5.5 was already a strong model for professional tasks: coding, research, document creation, spreadsheet work. Terra delivers that level of performance at a price point that makes high-volume use realistic for mid-sized businesses and agencies without enterprise budgets.
For marketers, Terra is where content workflows, campaign research, competitor auditing, and copy iteration live. It is capable enough to produce genuinely useful output on complex tasks without the cost overhead of Sol.
Pricing: $2.50 per million input tokens / $15 per million output tokens via API.
Availability: Currently in the same limited partner preview as Sol. Broad public release expected in the coming weeks alongside Sol and Luna.
GPT-5.6 Luna: Speed and Affordability for High-Volume Tasks
Luna sits at the bottom of the pricing tier but should not be dismissed. OpenAI describes it as a fast, affordable model with strong capability at the lowest cost in the GPT-5.6 family. It is built for tasks where latency matters and volume is high: think summarization pipelines, customer-facing chatbots, automated email drafting, or social media content at scale.
For agencies running high-output content programs or businesses with customer service automation, Luna fills the gap that heavier models cannot cost-effectively fill. You do not need Sol to summarize an inbox. You do not need Terra to draft a product description at scale. Luna handles that category of work reliably and cheaply.
Pricing: $1 per million input tokens / $6 per million output tokens via API.
Availability: Same limited preview as Sol and Terra. General availability expected in the coming weeks.
What This Means Right Now
If you are waiting to test these models, that is a reasonable position. The GPT-5.6 family is not publicly available yet. OpenAI is running a phased rollout with government-vetted partners first, following a request from the Trump administration ahead of a broader AI cybersecurity framework being developed through a June 2026 executive order. OpenAI has stated publicly that it does not want this kind of gated access to become the permanent default, and broad access is expected within weeks.
What you can do right now is assess where your current AI workflows are weakest. If you are cobbling together tasks manually between separate tools, a model like Sol’s multi-agent architecture changes that picture significantly. If you are looking for cost efficiency in content production, Terra and Luna will shift how you budget for AI-generated output once access opens.
The three-tier model structure OpenAI is rolling out is the direction the whole industry is heading. Knowing which tool to reach for, and when, is what separates teams that get ROI from AI from teams that are just paying for subscriptions.
