GPT-5.6 Terra
Chat model
Available on Ofrai — chat, images, and video in one place with shared credits.
Chat and video use credits from your Ofrai balance.
About 0.365696 credits per short prompt and 2.19418 credits per short reply (per ~1,000 characters of text).
Commercial use depends on the selected model and model terms.

About GPT-5.6 Terra on Ofrai
GPT-5.6 Terra is the long-context tier of OpenAI's current GPT-5.6 line, the one to use when the difficulty is how much material has to be held at once rather than how deep each step must go. On Ofrai it is listed in Text Studio and billed from the shared credit balance, and it costs more per request than Luna or Sol because it is built to work across very large inputs. Use it for document review, codebase-wide questions and multi-file synthesis where earlier context tends to get lost. Send it only the material that is actually relevant.
About 0.365696 credits per short prompt and 2.19418 credits per short reply (per ~1,000 characters of text).
GPT-5.6 Terra model facts
- Model ID
- gpt-5.6-terra
- Modality
- Chat model
- Released
- Ofrai pricing
- About 0.365696 credits per short prompt and 2.19418 credits per short reply (per ~1,000 characters of text).
- Web search
- Not available
Best for
- Reviewing a long document or a bundle of documents, such as a contract set, a manual or a policy corpus.
- Questions that span a whole codebase, where the answer depends on files far apart in the project.
- Synthesis across many sources, where each claim must be traced back to a specific place in the material.
- Long transcripts or case histories where an early detail changes the interpretation of everything after it.
Not the best fit when
- Short tasks, where Luna is faster and cheaper with an equally usable answer.
- Deep reasoning about one narrow problem, which is what Sol is built for.
- Sending enormous amounts of irrelevant text, since more input still costs more and dilutes the answer.
How it differs
Compared with GPT-5.6 Sol, GPT-5.6 Terra trades depth for breadth: use Terra when the challenge is holding a large amount of material accurately, and Sol when the challenge is working through a few hard steps.
Example prompts
Here is the full manual. Answer my question using only this document, quote the section number for every claim, and tell me explicitly when the document does not answer it.
I have attached twenty files from this project. Before answering, list the five files you actually used and say why the others were not relevant. Then answer the question and cite file and line for each finding.
Compare these four partner contracts clause by clause. Produce a table with one row per clause category, mark which party each obligation falls on, and highlight the three clauses most likely to cause a dispute.
Here is a long support history for one customer. Summarise the relationship so far, list the commitments made on each side with dates, and point out where our records contradict what was promised.
FAQ
- What is GPT-5.6 Terra used for?
- GPT-5.6 Terra is the long-context tier of OpenAI's current GPT-5.6 line. It is used when the hard part is the amount of material: reviewing a long document or document set, answering questions that span a whole codebase, synthesising many sources with traceable citations, and reading long transcripts where an early detail changes everything after it.
- How does GPT-5.6 Terra differ from GPT-5.6 Sol?
- Terra trades depth for breadth: it is built to hold a large amount of material accurately across a long input, while Sol is built to reason through a few hard steps on a smaller problem. If your difficulty is not finding the right file or passage, use Sol. If the difficulty is keeping everything in view at once, use Terra.
- Does GPT-5.6 Terra have a bigger context window?
- It is the tier to reach for when you need a lot of input at once, but treat the practical limit as what you can supply and get cited back correctly. Feeding a model irrelevant text does not make it more accurate; it costs more and makes answers vaguer. Curate the material and ask for citations so you can verify what was actually used.
- Is GPT-5.6 Terra better than Luna for long documents?
- For a genuinely long document with questions spread across it, yes, because Luna tends to lose earlier material. For a document you have not checked, start smaller: outline the document with Luna, then bring the relevant sections to Terra for detailed work. That combination is usually cheaper than sending everything at once.
- How is GPT-5.6 Terra billed on Ofrai?
- It is listed in Text Studio and charged against the shared credit balance along with your other Ofrai usage. Since input length drives the cost, the practical saving is in preparation: extract the sections that matter first, and reserve the full document for the questions that genuinely need it.
- When should I use a GPT-6 tier instead of Terra?
- Move up when long inputs stop being the problem and reliability becomes it: when the model summarises accurately but misses a nuance you know is there, or when combining a very long input with hard reasoning is needed at the same time. A GPT-6 flagship tier has more headroom for that combination than the current 5.6 family.
Related models
More models in the same modality on Ofrai.
