OpenAI's GPT-5.6 family splits into three tiers — Sol, Terra, and Luna — each priced and tuned for a different job, from heavy-duty coding agents to cheap, high-volume automation.
On July 9, 2026, OpenAI rolled out GPT-5.6 — not as a single model, but as a family of three: Sol, Terra, and Luna. Each tier targets a different slice of real-world usage, letting teams pick capability and cost independently instead of defaulting to "the biggest model" for every task. The rollout followed a limited preview that began on June 26, 2026, per OpenAI's own GPT-5.6 announcement.
Three tiers, three jobs
The family is ordered by capability and cost, and each model is aimed at a distinct workload:
- Sol — the flagship. Built for the hardest problems: complex coding, long-running agentic tasks, cybersecurity research, and scientific work. OpenAI's preview post for Sol positions it as their new state-of-the-art model for intelligence and efficiency together, not just raw benchmark scores.
- Terra — the balanced option. Aimed at high-volume business tasks like customer support, internal tooling, and document analysis. OpenAI says it matches GPT-5.5-level performance at roughly half the cost.
- Luna — the fast, cheap tier. Built for everyday work: summarization, drafting, and routine automation where latency and cost matter more than frontier reasoning.
Pricing
All three are priced per 1M tokens:
- Sol: $5 input / $30 output
- Terra: $2.50 input / $15 output
- Luna: $1 input / $6 output
All three share the same underlying specs: a 1 million token context window, 128,000 maximum output tokens, and a February 16, 2026 knowledge cutoff.
For context against the competing frontier tier, Anthropic's Claude Fable 5 — its most capable widely released model — is priced at $10 input / $50 output per 1M tokens. That puts GPT-5.6 Sol at roughly half the per-token cost of Fable 5, though the two aren't a straight swap: Fable 5 targets the most demanding long-horizon agentic and reasoning work specifically, while Sol is OpenAI's do-everything flagship across coding, cybersecurity, and science.
| Model | Input ($/1M tokens) | Output ($/1M tokens) |
|---|---|---|
| GPT-5.6 Luna | $1.00 | $6.00 |
| GPT-5.6 Terra | $2.50 | $15.00 |
| GPT-5.6 Sol | $5.00 | $30.00 |
| Claude Fable 5 | $10.00 | $50.00 |
What's new under the hood
Beyond raw capability, GPT-5.6 ships with a few API-level changes worth knowing about if you're building on it:
- Max reasoning effort — a new setting for Sol that pushes reasoning depth further for the hardest tasks.
- Ultra mode — coordinates multiple subagents in parallel (four, by default) to accelerate complex workflows. OpenAI reports it lifts Terminal-Bench 2.1 performance from 88.8% to 91.9%.
- Programmatic tool calling — lets the model write and run JavaScript directly in an isolated V8 runtime (no network access) inside the Responses API, cutting out round-trips for tool orchestration. Early adopters reported token reductions between roughly 38% and 63%.
- Prompt cache breakpoints — explicit control over where caching kicks in, with a 30-minute minimum cache life.
Where to get it
Sol is available now in ChatGPT for Plus, Pro, Business, and Enterprise users at medium-and-above reasoning effort. All three models — Sol, Terra, and Luna — are available via the OpenAI API with multi-agent support in the Responses API, and OpenAI's Codex. GitHub also shipped support the same day, per the GitHub Copilot changelog, making all three tiers selectable inside Copilot.
Our take
The headline here isn't really "a smarter model" — it's the explicit tiering. Sol, Terra, and Luna make the cost/capability tradeoff a first-class decision instead of an afterthought, which matters a lot once you're running agents at volume rather than answering one-off prompts. For teams already budgeting API spend by workload, that's arguably more useful than another jump on a benchmark leaderboard — and with Sol landing well under Fable 5 on a per-token basis, the pricing gap is now part of that decision too.
Source: OpenAI — GPT-5.6: Frontier intelligence that scales with your ambition
