Check GPT-5.6 Access Before Moving Agent Workflows
OpenAI has made GPT-5.6 generally available across ChatGPT, Codex, and the API, with Sol, Terra, and Luna tiers, new multi-agent options, and stricter cyber safeguards.
OpenAI has confirmed general availability for GPT-5.6 on July 9, 2026. Confidence level: confirmed for the model family, pricing, ChatGPT, Codex, and API availability; account rollout may still take up to 24 hours. The practical question is which tier your workflow gets: Sol for highest capability, Terra for balanced work, or Luna for faster lower-cost tasks.
What changed
OpenAI moved GPT-5.6 from preview to general availability and split the family into three durable tiers. Sol is the flagship model, Terra is the balanced model, and Luna is the faster, lower-cost model. OpenAI says the family is available across ChatGPT, Codex, and the OpenAI API, with the rollout moving toward full availability over the next 24 hours.
The launch also adds two workflow controls that matter for agents. max gives GPT-5.6 more time to reason and revise, while ultra coordinates multiple agents in parallel for harder work. In the API, OpenAI says Programmatic Tool Calling and a multi-agent beta can help developers build similar long-running workflows.
| Model | Best fit | Access | Price per 1M tokens | Caveat |
|---|---|---|---|---|
| GPT-5.6 Sol | Highest-capability coding, research, design, and cyber review | ChatGPT paid plans, Codex, API | $5 input / $30 output | Pro and Enterprise get the strongest Chat options |
| GPT-5.6 Terra | Balanced everyday agent and knowledge work | ChatGPT Work, Codex, API | $2.50 input / $15 output | Verify quality against your own GPT-5.5 baseline |
| GPT-5.6 Luna | Fast, lower-cost repeated tasks | ChatGPT Work, Codex, API | $1 input / $6 output | Lower price does not mean every task should route here |
Sol ultra | Parallel agent work on complex tasks | ChatGPT Work Pro/Enterprise, Codex Plus and higher | Higher token use | Use only where time-to-result beats the extra spend |
Why this is early
TestingCatalog flagged GPT-5.6 signals before the official launch, including the Sol, Terra, and Luna naming and Codex-related rollout clues. That made it useful as an early-watch source, but the publishable claim now rests on OpenAI's own July 9 launch post and system card.
The early part is operational, not factual. OpenAI says availability starts today and rolls out globally, but users still need to check the exact model picker, API account, Codex environment, workspace tier, region, and admin policy they use. A launch page is not the same as visible access in every account.
Key takeaways
- GPT-5.6 is now a confirmed general-availability release, not only a preview.
- Sol, Terra, and Luna give teams clearer routing choices by capability and cost.
- ChatGPT, Codex, and API access are included, but account rollout may lag.
- Programmatic Tool Calling and multi-agent beta support make the release more important for agent builders than a plain chatbot upgrade.
- Cyber and biology use cases need extra review because OpenAI describes stronger capability and stricter safeguards in the system card.
Availability and access
OpenAI says Plus, Pro, Business, and Enterprise users can access GPT-5.6 Sol through medium and higher effort settings in Chat. Pro and Enterprise users can select GPT-5.6 Sol Pro for complex work. In ChatGPT Work and Codex, Free and Go users get Terra, while Plus, Pro, Business, and Enterprise users can choose Sol, Terra, and Luna where available.
For developers, Sol, Terra, and Luna are available through the OpenAI API. OpenAI lists Programmatic Tool Calling in the Responses API as Zero Data Retention compatible, and describes multi-agent as an initial beta. Pricing is public, but real-world cost still depends on cache behavior, tool calls, retries, and how often teams use max or ultra.
Practical LinkLoot angle
GPT-5.6 is most relevant where agent success rates, tool use, and artifact quality change the economics of a workflow. Good first tests are repository repair, frontend generation, code review, spreadsheet or deck creation, multi-step research, and defensive security triage in authorized environments.
The routing decision matters more than the headline. Keep GPT-5.5, Claude, Gemini, or open-weight fallbacks live while you test Sol, Terra, and Luna on real tasks. For agent stack planning, pair this rollout with LinkLoot's guide to AI agent tools.
What to verify before you act
- Confirm GPT-5.6 appears in your ChatGPT, Codex, API, or workspace model picker.
- Check whether your plan exposes Sol, Terra, Luna,
max, andultra, not just one tier. - Recalculate cost with input, output, cache write, cache read, tool-call, and retry behavior.
- Read the GPT-5.6 system card before cyber, biology, or other dual-use workflows.
- Run side-by-side tests against your current model on real prompts, not vendor demos.
Source check
Confirmed by OpenAI: GPT-5.6 is generally available starting July 9, 2026; the family includes Sol, Terra, and Luna; ChatGPT, Codex, and API access are included; API pricing is $5/$30 for Sol, $2.50/$15 for Terra, and $1/$6 for Luna per 1M input/output tokens; GPT-5.6 includes updated safeguards described in the system card.
Independent context: Axios reported the broad release and the ChatGPT Work launch context. TestingCatalog served as the early signal source before the official release. LinkLoot will treat API model IDs, workspace admin notes, or safety-card revisions as update triggers, not as a reason to blur today's confirmed launch with earlier rumors.
Yes. OpenAI announced GPT-5.6 general availability on July 9, 2026, with rollout across ChatGPT, Codex, and the API.