OpenAI Previews GPT-5.6: Sol, Terra, and Luna in Limited Preview
OpenAI announced a limited preview of the GPT-5.6 family: Sol, a next-generation frontier flagship OpenAI calls a step function better than GPT-5.5; Terra, a balanced model competitive with GPT-5.5 at 2x lower cost; and Luna, its most cost-efficient model. Access starts with trusted partners in Codex and the API.
OpenAI has announced a limited preview of GPT-5.6, a new model family with three members: GPT-5.6 Sol, a next-generation frontier flagship; GPT-5.6 Terra, a balanced model for efficient everyday work; and GPT-5.6 Luna, a fast and affordable model for high-volume work. OpenAI calls Sol a step function better than GPT-5.5 and its most capable model yet for cybersecurity. For now, access is restricted to a small group of trusted partners in Codex and the API. (OpenAI on X, OpenAI)
Key Takeaways
- Three new models in limited preview: Sol (next-generation frontier flagship), Terra (balanced, for efficient everyday work), and Luna (fast and affordable, for high-volume work).
- Sol is the new flagship: OpenAI describes it as a step function better than GPT-5.5.
- Terra trades the same performance for cost: OpenAI says Terra is competitive with GPT-5.5 at 2x lower cost.
- Luna is the cheapest tier: OpenAI calls it its most cost-efficient model, delivering strong capability at its lowest cost.
- Cybersecurity is the headline strength: OpenAI calls Sol its most capable model yet for cybersecurity, saying it shifts the performance-efficiency frontier for long-horizon security tasks including vulnerability research and exploitation.
- New state of the art on Terminal-Bench 2.1: OpenAI says Sol sets a new state of the art on this command-line agentic benchmark.
- Restricted access for now: OpenAI says it is starting with a limited preview among a small group of trusted partners in Codex and the API, at the request of the U.S. government, with general availability planned in the coming weeks.
What OpenAI Announced
OpenAI introduced the GPT-5.6 family as a limited preview, positioning each model at a different point on the capability-and-cost curve:
- GPT-5.6 Sol is the next-generation frontier flagship. OpenAI says it is a step function better than GPT-5.5.
- GPT-5.6 Terra is the balanced option for efficient, everyday work. OpenAI says Terra delivers performance competitive with GPT-5.5 at 2x lower cost.
- GPT-5.6 Luna is the fast, affordable model for high-volume work. OpenAI calls it its most cost-efficient model, delivering strong capability at its lowest cost.
The framing is a three-tier lineup: a new top-end flagship, a mid tier that matches the prior flagship for less money, and a low-cost tier for high-volume workloads.
Cybersecurity and Agentic Coding
The strongest claims OpenAI makes are about Sol on security and command-line work. OpenAI calls GPT-5.6 Sol its most capable model yet for cybersecurity, and says it sets a new state of the art on Terminal-Bench 2.1, the agentic command-line benchmark.
On security, OpenAI says Sol shifts the performance-efficiency frontier for long-horizon security tasks including vulnerability research and exploitation. On agentic coding, OpenAI says Sol sets a new state of the art on Terminal-Bench 2.1, a benchmark that tests complex command-line workflows requiring planning, iteration, and tool coordination. OpenAI did not attach a specific Terminal-Bench 2.1 score to the announcement, so we are not citing one.
On safety, OpenAI says Sol launches with its most robust safety stack yet. Per OpenAI, it strengthened real-time protections against high-risk cyber activity and repeated misuse, then spent weeks hardening the system with human red teaming and over 700,000 A100-equivalent GPU hours of automated testing. The emphasis on cyber safeguards tracks with the availability decision below.
Availability: Limited Preview First
OpenAI says it believes in broad access and plans to make GPT-5.6 Sol, Terra, and Luna generally available in the coming weeks. For now, at the request of the U.S. government, OpenAI is starting with a limited preview among a small group of trusted partners in Codex and the API.
The practical near-term takeaway is access, not benchmarks: the GPT-5.6 models are not broadly available yet. If you rely on Codex or the OpenAI API, GPT-5.6 is the next flagship step, but most developers will wait for the wider rollout OpenAI says is coming.
This follows the April 23 GPT-5.5 launch that introduced the prior flagship, and the more recent GPT-5.5 Instant update in ChatGPT.
What We Are (and Aren't) Claiming
This piece is built on OpenAI's official announcement on X and the linked OpenAI preview page. We are claiming only what OpenAI stated:
- A limited preview of the GPT-5.6 family (Sol, Terra, Luna), with the positioning OpenAI gave each model.
- Sol is the flagship and a step function better than GPT-5.5; Terra is competitive with GPT-5.5 at 2x lower cost; Luna is the most cost-efficient.
- Sol is OpenAI's most capable model yet for cybersecurity, sets a new state of the art on Terminal-Bench 2.1, and ships with OpenAI's most robust safety stack yet (human red teaming plus over 700,000 A100-equivalent GPU hours of automated testing).
- Limited preview in Codex and the API for trusted partners now, at the request of the U.S. government, with general availability planned in the coming weeks.
We are not adding specific benchmark scores, pricing, model-card details, context windows, or availability specifics beyond the above, because OpenAI did not publish those alongside this announcement.
Sources
- OpenAI announcement on X (Jun 26, 2026) -- https://x.com/OpenAI/status/2070555272230384038
- OpenAI: Previewing GPT-5.6 Sol -- https://openai.com/index/previewing-gpt-5-6-sol/
Read next
Keep building the workspace playbookGPT-5.5 Is Here: State-of-the-Art Agentic Coding, 1M Context, and a New Pro Tier
OpenAI launched GPT-5.5 on April 23, 2026 -- its smartest model yet, with state-of-the-art scores on Terminal-Bench 2.0 (82.7%), GDPval (84.9%), and OSWorld-Verified (78.7%), GPT-5.4 per-token latency, and a new GPT-5.5 Pro tier for harder work. Available in ChatGPT and Codex today, with API at $5/M input and $30/M output coming soon.
OpenAI Ships a New GPT-5.5 Instant in ChatGPT: Better Intent, Constraints, and Shopping/Local Recs
OpenAI says a new version of GPT-5.5 Instant is rolling out in ChatGPT with better intent understanding, more reliable handling of complex constraints, and improved shopping and local recommendations. Per OpenAI, it reaches paid users first and free users the next day. GPT-5.5 Instant is ChatGPT's default for logged-in users.
Frequently Asked Questions
What is the GPT-5.6 family?
Per OpenAI, GPT-5.6 is a new model family in limited preview with three members: GPT-5.6 Sol, a next-generation frontier flagship; GPT-5.6 Terra, a balanced model for efficient everyday work; and GPT-5.6 Luna, a fast and affordable model for high-volume work. OpenAI calls Sol a step function better than GPT-5.5.
Who can use GPT-5.6 right now?
OpenAI says it is starting with a limited preview among a small group of trusted partners in Codex and the API, at the request of the U.S. government. OpenAI says it plans to make Sol, Terra, and Luna generally available in the coming weeks.
How do GPT-5.6 Terra and Luna differ from Sol?
Per OpenAI, Sol is the new flagship and a step function better than GPT-5.5. Terra delivers performance competitive with GPT-5.5 at 2x lower cost. Luna is OpenAI's most cost-efficient model, delivering strong capability at its lowest cost.
What does OpenAI say about GPT-5.6 Sol and cybersecurity?
OpenAI calls GPT-5.6 Sol its most capable model yet for cybersecurity, saying it shifts the performance-efficiency frontier for long-horizon security tasks including vulnerability research and exploitation. OpenAI says Sol ships with its most robust safety stack yet.