GPT-5.6 Sol vs. Luna vs. Terra: How to Choose Among OpenAI's Three Tiers

Of the three GPT-5.6 versions, Terra is the default choice for general work. Switch to Sol for long, multi-step Agent tasks. Luna only fits short-text, high-vol

Of the three GPT-5.6 versions, Terra is the default choice for general work. Switch to Sol for long, multi-step Agent tasks. Luna only fits short-text, high-volume pipelines. Two numbers support this. First, Luna scores just 41.3% on a long-context recall test, while Terra scores 89.6%. Second, on long-horizon coding tasks, Sol produces less than half as many output tokens per task as Terra. Its per-token price is higher, but the actual cost per task is not necessarily higher. What Sol, Terra, and Luna Are: OpenAI's New Naming Scheme GPT-5.6 is a model family OpenAI officially released on July 9, 2026, in three versions: Sol, Terra, and Luna. According to Wikipedia's GPT-5.6 entry , it first opened as a preview to a small group of partners on June 26, 2026, and became generally available two weeks later. The new naming scheme separates "generation" from "tier." The number 5.6 is the model generation. Sol (sun), Terra (earth), and Luna (moon) are three long-term, fixed capability tiers, and each tier can be upgraded on its own. You may later see Terra updated while Sol is not, so choose a model by its tier, not just its version number. Sol : The most capable tier, for frontier reasoning, long-horizon Agent tasks, security research, and computer use. It also offers a max reasoning effort and an ultra mode that uses subagents. Terra : The mid tier that balances capability and cost. OpenAI says it performs on par with GPT-5.5 at roughly half the cost. Luna : The lightweight, low-latency, low-cost tier, for high-volume, latency-sensitive, or budget-limited work. The three versions share most specs and differ in capability and price. According to Codersera's spec table , all three have a 1,050,000-token context window, up to 128,000 output tokens, and a knowledge cutoff of February 16, 2026. Reasoning effort can be set from none to max , with medium as the default. Price Comparison: The Gap Widened After the July 30 Price Cut On July 30, 2026, OpenAI cut prices: Luna by 8

Related Guidebooks

Reviewed and verified by FeiYueh · Last verified 2026-10-04. Independently maintained — not AI-generated boilerplate.

← Back to Blog