GPT-5.6 Terra sits in the middle of OpenAI's GPT-5.6 lineup as the balanced option between the flagship Sol tier and the more economical Luna tier. Independent routing directories describe it as tuned for everyday coding, general reasoning, and agentic workflows where teams want strong capability without paying the top-of-line Sol rate, with published positioning suggesting roughly half the cost of Sol for comparable workloads. Within the GPT-5.6 series this places Terra as the pragmatic default for production agents, developer assistants, and document-heavy pipelines that need dependable reasoning alongside prompt caching and tool use.

Practically, Terra behaves like a long-context workhorse: a roughly one-million-token window with up to the cataloged API limit output tokens, and a long-context surcharge that doubles input and applies a 1.5x multiplier to output once requests cross a the listed price threshold. Routing layers expose the same model through multiple hosts with different latency, throughput, and uptime trade-offs, and pricing tiers include reduced rates for cached-input reads, which can lower effective spend substantially when repeated that quick-info value is reused. Coding evaluations published alongside the release show Terra posting competitive scores on SWE-bench Pro and Terminal-Bench 2.1, reinforcing its fit for terminal-style agent coding and OpenAI's Responses API workflows rather than purely conversational use.