7/15/2026, 1:01:36 PM · foundation-models

OpenAI Releases GPT-5.6 Family — Sol, Terra, and Luna — Following Staged U.S. Government Review

OpenAI's three-tier GPT-5.6 model family entered a government-coordinated limited preview before broader rollout, marking a new pattern for frontier Large Language Model releases under a June 2026 White House executive order.

Overview

OpenAI on July 9, 2026, began a limited preview of GPT-5.6, a three-model family that introduces a new naming convention for its frontier Large Language Model (LLM) lineup. <cite index="1-1">The series comprises Sol, the flagship model; Terra, a balanced model for everyday work; and Luna, a fast and affordable model.</cite> <cite index="1-13">In this new naming system introduced with GPT-5.6, the number identifies a model's generation, while Sol, Terra, and Luna identify durable capability tiers that can advance on their own cadence.</cite>

Staged Government Rollout

The release did not proceed as a standard public launch. <cite index="1-7,1-8">As part of OpenAI's ongoing engagement with the U.S. government, the company previewed its plans and the models' capabilities ahead of the launch. At the government's request, OpenAI began with a limited preview for a small group of trusted partners whose participation was shared with the government.</cite> <cite index="3-5,3-6">The models were made available initially to approximately 20 organizations, with a general release planned for "the coming weeks."</cite>

This staged approach reflects a broader policy shift. <cite index="12-6">On June 2, 2026, President Trump issued an executive order (EO) directing U.S. government agencies to accelerate artificial intelligence (AI)-enabled cybersecurity initiatives and design a voluntary framework for engagement with developers of frontier AI models before broader release.</cite> <cite index="14-3">For the private sector, the EO establishes a voluntary process for AI model developers to submit AI models for federal review prior to broader release, and provide critical infrastructure entities with early access to these models to strengthen cybersecurity protections.</cite> <cite index="10-24">Companies developing advanced AI models are asked to participate in a voluntary review process that gives the government a window of 30 days to examine new systems before such models are released to "other trusted partners."</cite>

<cite index="10-15">Previous agreements between the National Institute of Standards and Technology (NIST) and AI companies Anthropic and OpenAI had already established a framework for NIST to "receive access to major new models from each company prior to and following their public release."</cite> <cite index="13-5">On May 5, 2026, the Center for AI Standards and Innovation (CAISI), housed at the Department of Commerce's NIST, announced agreements to conduct pre-deployment evaluations and targeted research on frontier AI systems from Google DeepMind, Microsoft, and xAI.</cite>

Capability Tiers and Pricing

<cite index="6-2">Sol is built for frontier reasoning and long-horizon agentic work; Terra is a balanced everyday model with GPT-5.5-competitive performance at 2x lower cost; and Luna is the fastest, most affordable member of the family.</cite> <cite index="1-15">GPT-5.6 is priced per one million tokens: Sol at $5 input / $30 output; Terra at $2.50 input / $15 output; and Luna at $1 input / $6 output.</cite> <cite index="8-5">All three models share a February 16, 2026 knowledge cutoff, a one-million-token context window, and 128,000 maximum output tokens.</cite>

On benchmarks, <cite index="7-3,7-4">GPT-5.6 Sol set a new high of 53.6 on Agents' Last Exam — an evaluation of long-running professional workflows across 55 fields — eclipsing Claude Fable 5 by 13.1 points, with the gap holding even at medium reasoning at roughly one-quarter the estimated cost.</cite> OpenAI also noted a Cerebras partnership: <cite index="1-18">GPT-5.6 Sol is launching on Cerebras at up to 750 tokens per second in July, bringing frontier intelligence to customers at unprecedented speed.</cite>

Safety Classification and Governance Implications

<cite index="3-19,3-20">OpenAI's system card classifies all three GPT-5.6 models — not just Sol — at its "High" risk level for both cyber and biological/chemical capability, meaning even the cheaper Terra and Luna tiers may carry new governance obligations for companies using them in security, life sciences, or other sensitive workflows.</cite>

<cite index="11-11">Although voluntary, the executive order is expected to shape industry norms and may serve as the basis for subsequent contractual, procurement, or regulatory requirements.</cite> <cite index="16-17">While some commentators have welcomed this measured approach, others have warned it could prove a stepping stone to more prescriptive regulation.</cite>

General Availability

<cite index="7-1,7-2">GPT-5.6 became available across ChatGPT, Codex, and the OpenAI Application Programming Interface (API), with the rollout starting globally and continuing gradually toward full availability over 24 hours.</cite> <cite index="7-24,7-25">Plus, Pro, Business, and Enterprise users access GPT-5.6 Sol through medium and higher effort settings, while Pro and Enterprise users can also select GPT-5.6 Sol Pro for the highest-quality results on complex tasks.</cite>

Cross-references

Sources

  1. [1]
    Previewing GPT-5.6 Sol: a next-generation model | OpenAI
  2. [2]
    GPT-5.6 in ChatGPT | OpenAI Help Center
  3. [3]
    OpenAI unveils GPT-5.6 Sol, Terra and Luna models — but only accessible to limited preview partners for now, per US Gov | VentureBeat
  4. [4]
    GPT-5.6 now widely available: Sol, Terra, and Luna launched - Techzine Global
  5. [5]
    [AINews] OpenAI launches GPT 5.6 Sol/Terra/Luna, Codex becomes ChatGPT superapp
  6. [6]
    Introducing GPT-5.6 series: Sol, Terra and Luna. Coming July 9 10am PT - Announcements - OpenAI Developer Community
  7. [7]
    GPT-5.6: Frontier intelligence that scales with your ambition | OpenAI
  8. [8]
    The new GPT-5.6 family: Luna, Terra, Sol
  9. [9]
    OpenAI Launches GPT-5.6 With Sol, Terra, And Luna Models - Dataconomy
  10. [10]
    Controlling Advanced Artificial Intelligence: Executive Order 14409 Explained | Congress.gov | Library of Congress
  11. [11]
    New Executive Order Addressing Early Government Access to Frontier AI Models
  12. [12]
    New AI Executive Order Calls for Frontier Model Security, Early Government Access and AI-Enabled Cyber Defense | Skadden, Arps, Slate, Meagher & Flom LLP
  13. [13]
    Pre-Deployment AI Evaluation Moves From China’s Model To Washington
  14. [14]
    New AI Executive Order Addresses Frontier Models and Cybersecurity Vulnerabilities: Wiley
  15. [15]
    Letter to White House: Guidance on Oversight of AI Models
  16. [16]
    New AI Executive Order Calls for Frontier Model Security, Early Government Access and AI-Enabled Cyber Defense - Lexology
  17. [17]
    US Government Pushes Pre-Release AI Model Reviews
  18. [18]
    US order seeks 30-day frontier model access before public release