| Model ID | Best for |
|---|---|
gpt-5.6-sol | Complex reasoning, large codebases, multi-step agent tasks, and long-running coding workflows |
gpt-5.6-terra | Everyday coding, interactive development, common agent tasks, and balanced capability and cost |
gpt-5.6-luna | Lighter tasks where speed and cost matter, such as batch assistance, quick Q&A, simple rewrites, and small tool calls |
Highlights
- Three model tiers: OpenAI positions GPT-5.6 Sol, Terra, and Luna for high-reasoning work, balanced everyday tasks, and lighter high-volume tasks
- Better fit for coding agents: OpenAI highlights improvements across coding-agent, terminal workflow, and tool-coordination tasks
- Ready for external tools: OpenAI Help Center confirms Sol, Terra, and Luna are available through the OpenAI API; in LLMEasy, use the GPT Key group and
https://www.llmeasy.ru/v1 - Choose by workload: Start with
gpt-5.6-solfor complex work,gpt-5.6-terrafor daily development, andgpt-5.6-lunafor lighter or cost-sensitive requests
Setup guidance
For Codex or other OpenAI-compatible tools, follow the matching setup guide:- Base URL:
https://www.llmeasy.ru/v1 - API Key group: choose the GPT Key group
- Model name: copy the current model ID from the model plaza
Do not hardcode a model name that is not shown in the model plaza. Model availability, pricing, and routing status follow the live model plaza.
How to choose
If you are not sure where to start, choose by task complexity:- Use
gpt-5.6-solfor large codebase understanding, cross-file edits, test repair, long debugging loops, and agent tasks that need stronger reasoning - Use
gpt-5.6-terrafor everyday coding, code explanation, routine refactors, documentation, and most interactive development tasks - Use
gpt-5.6-lunafor lightweight Q&A, batch assistance, simple text processing, low-latency tool calls, and high-volume cost-sensitive requests

