Tool support requires the appropriate API integration; a supported tool is not automatically active in every chat.
Supported API features and tools
Streaming responses and prompt caching
Function calling and structured outputs
Web search and file search
Code interpreter and hosted shell
Computer use and image generation tools
MCP, tool search, skills, and apply patch
Before you choose
No native audio or video support.
Image generation requires a tool; native output is text.
Fine-tuning is not supported.
Choose the reasoning effort
nonelowmedium · defaulthighxhighmax
Use medium as a baseline on tasks with a clear review rubric. Try none or low for narrow transformations, and compare higher effort when the task has dependencies or conflicting evidence. Measure the time you spend correcting answers, not just the model's response time. Extra reasoning is not a substitute for supplying missing information.
Unsupported settings: minimal.
Use the Responses API for built-in tools and multi-turn reasoning workflows. Chat Completions and Batch are also supported.
The model documentation lists a maximum input of 922,000 tokens, separately from its 1,050,000-token context window and output limit.
Use gpt-5.6-terra explicitly; the gpt-5.6 family alias points to Sol, not Terra. API rate limits depend on your usage tier.
Put it to work
Start with a more useful prompt.
Original examples from EZ Ai Assist. Adapt these to your task and the features available in your workspace.
Workflow 01
Turn notes into a working brief
Supply the meeting notes and constraints so decisions stay separate from suggestions.
Turn these meeting notes into a working brief for the project described below. Separate agreed decisions, proposed ideas, open questions, and next actions. For each action, include the owner and deadline only if the notes state them; otherwise mark them as unassigned. Cite the note or speaker supporting each decision. End with the three questions that most affect our next step. Do not invent commitments or send messages.
Workflow 02
Review a focused bug fix
Keep the review bounded by a reproducible problem and the smallest relevant code change.
Review this proposed fix using the reproduction steps, expected behavior, and diff below. Explain whether the changed code addresses the reported cause, identify edge cases it could break, and cite the relevant lines. Suggest the smallest adjustment if needed. Give three targeted regression tests with inputs and expected outcomes. Distinguish tests you recommend from tests actually run. Do not refactor unrelated code or apply changes.
Workflow 03
Draft a grounded customer reply
Combine an inquiry with approved product information without inventing policies or promises.
Draft a concise reply to this customer inquiry using only the approved product notes and policies below. Acknowledge the question, provide the supported answer in plain language, and suggest one practical next step. List the source for each product or policy claim separately from the draft. If the material does not answer part of the inquiry, flag it for a human rather than guessing. Do not promise a refund, deadline, or feature that is not documented, and do not send the reply.
Developer reference
OpenAI API pricing
These are OpenAI API reference prices, not EZ Ai Assist subscription prices.
Above 272,000 input tokens, long-context rates apply to the entire request, not just the excess tokens.
Cache writes cost 1.25× the uncached input rate; cached reads use the separate cached-input rate.
This table covers Standard processing. Other service tiers and tool usage can change the final bill. Confirm current rates before estimating API spend.
Common questions
A few things worth knowing.
When should I choose Terra instead of Luna or Sol?
OpenAI positions Terra between cost-focused Luna and more capable Sol, roughly in the earlier mini tier. Build a small evaluation set from your own work: a routine task, an ambiguous one, and a difficult edge case. Compare useful answers, corrections, latency, and total cost. Treat model positioning as a starting point, not a guarantee of equal quality across tasks.
What should I include in an everyday work prompt?
State the outcome, supply the relevant material, and describe the output format. Mark which constraints are firm and which decisions remain open. Give an example of an acceptable answer when formatting matters. A bounded request is easier to evaluate than asking for a general improvement with no success criteria.
How should I handle a large source packet?
Give each source a stable label, identify the question to answer, and remove unrelated material. Request citations and inspect them against the originals. Keep room for the response and account for both input and output limits. More context is capacity, not a promise that every detail will be recalled or weighed correctly.
Can Terra help interpret a screenshot?
Image input is supported. Pair a screenshot with the task, relevant text, and any behavior the image cannot show. Ask for observations separately from assumptions. Verify suggested changes in the working interface; a screenshot alone cannot establish keyboard access, functional correctness, or behavior at other screen sizes.
How do I know whether a tool actually ran?
Check the integration's tool records or returned results rather than relying on a sentence in the answer. Supported tools must be enabled and given the appropriate permissions. Keep consequential actions behind review, and ask the model to distinguish evidence obtained from a tool from work it is only proposing.
Does a lower API rate always mean a cheaper workflow?
No. Compare the whole task, including prompt length, output length, reasoning, retries, tool fees, and review time. Use the same inputs and acceptance criteria when comparing models. For repetitive work, measure a sample before projecting volume; a response that needs repeated correction may cost more overall.
Are these rates my EZ Ai Assist subscription price?
No. The table describes direct OpenAI API usage. EZ Ai Assist subscription plans, model access, and available controls are separate. Use our pricing page for subscriptions and check the app before relying on a particular API feature. The example prompts are editorial starting points, not guaranteed outcomes.
Check the source
Official documentation
Specifications and API prices checked on . Example prompts and workflow advice are editorial guidance from EZ Ai Assist.