OpenAI launched a limited preview of its GPT-5.6 model family on June 26, 2026, introducing three distinct tiers for different enterprise workloads: Sol, Terra, and Luna. Access is currently restricted to roughly 20 trusted partner organizations while OpenAI completes broader safety validation, with general availability expected within weeks.

Three Models, One Architecture

The GPT-5.6 family replaces the previous GPT-5.5 generation with a cleaner three-tier structure:

Sol is OpenAI’s strongest model to date. It targets demanding agentic workflows — multi-step coding tasks, scientific reasoning, and cybersecurity analysis — and carries the most robust safety stack OpenAI has shipped. Pricing starts at $5 per million input tokens and $30 per million output tokens.

Terra is designed for everyday enterprise use. It delivers performance comparable to GPT-5.5 but at half the cost: $2.50 per million input tokens and $15 per million output tokens. For organizations running large volumes of AI-assisted workflows, that cost delta compounds quickly.

Luna is the fastest and most affordable option at $1 per million input tokens and $6 per million output tokens. It is positioned for high-frequency, lower-complexity tasks — call routing, data extraction, document classification — where speed and cost matter more than peak capability.

All three models are available through the OpenAI API and Codex during the preview period.

The Government Briefing Dimension

OpenAI shared the models and release plans with the U.S. government before making them available to partners — a deliberate step that reflects the regulatory environment taking shape in 2026. The move aligns with the direction both major AI labs and policymakers are traveling: increased transparency and pre-release coordination for frontier systems.

This matters for enterprise procurement teams. Organizations that need to justify AI tool choices to regulators or internal governance boards can point to a model family that was validated before reaching the market.

What This Means for Business

Lower costs for AI-powered operations. Terra’s 2x cost reduction over GPT-5.5, combined with Luna’s entry-level pricing, changes the economics of running AI agents at scale. If you are processing thousands of documents, handling high volumes of customer enquiries, or running continuous data pipelines, the difference between $15 and $30 per million tokens adds up fast.

Better agentic performance in Sol. The improvements OpenAI highlights in Sol — coding, biology, and cybersecurity — signal where the frontier is moving: agents that can execute complex, multi-step tasks with greater reliability. For businesses building AI-assisted workflows, Sol’s agentic improvements reduce the failure rate on longer-horizon tasks.

A cleaner model selection decision. The Sol / Terra / Luna structure is more intuitive than the GPT-5 to GPT-5.5 naming. Businesses can match workload type to model tier more easily: Sol for high-stakes autonomous tasks, Terra for general-purpose AI work, Luna for volume processing.

Near-term availability. The preview period is narrow by design. General availability is expected within weeks, meaning procurement and IT teams should start evaluating now rather than waiting for the official launch.

The Broader Picture

GPT-5.6 arrives alongside Anthropic’s continuing rollout of the Claude Opus 4.x series and Microsoft’s MAI model family, all competing for enterprise AI budget. The competitive pressure is benefiting customers — Terra’s per-token cost at GPT-5.5 capability is a direct response to pricing pressure from alternatives.

For business owners and operations leaders, the story is simple: AI model costs are falling while capability is rising. The practical question is not whether to use frontier AI in your workflows — it is which tier fits each task, and whether your team has the infrastructure to take advantage of it.

If you’re building with Claude or Codex right now, grab the free Working With Claude field guide. Thirty-two pages on the full ecosystem, Claude Code in depth, and how to roll agents out properly. Get the free guide.