See how ChatGPT, Perplexity and Google AI Overviews describe you today

Free AI Visibility Audit
Surge45°
OpenAI model°

GPT-5.6 Terra

GPT-5.6 Terra is the mid-tier model in OpenAI's GPT-5.6 series, sitting between the flagship Sol tier and the cheaper Luna tier, and built for everyday coding, reasoning and agentic work.

OpenAIVerified 24/09/2026Released 09/07/2026

01 / decision

Decision snapshot

Each figure sits next to the middle of the field, so it can be read as dear or cheap, wide or narrow, rather than floating on its own. Compared against the 151 models we track, not against an absolute standard.

Capability

Not measured

No published benchmark scores yet

Input / 1M tokens

$2.00

field median $0.43

Output / 1M tokens

$12.00

field median $1.75

Context

1,050K

field median 524K

02 / overview

What GPT-5.6 Terra is, and when to reach for it

Four questions, answered separately, because somebody arrives at one of them rather than at the top of the page.

What it is

The job it was built for, and the job it is not for.

GPT-5.6 Terra is a general-purpose model aimed at the bulk of production work rather than the hardest problems or the cheapest ones. It handles text, images and file input, so it can read a document or a screenshot as part of a task instead of needing a separate extraction step. It is not the tier to reach for when a task genuinely needs the top of the range, and it is not the tier to reach for on high-volume, low-value calls either.

When it arrived, and when to use it

Where it sits in its line, and when a sibling is the better pick.

GPT-5.6 Terra first appeared in July 2026 as the balanced member of the GPT-5.6 line. It is the right pick when a workload is varied enough that you cannot predict which requests will be hard, and you want one model that copes with all of them. If the work is uniformly simple, Luna is the better economics, and if it is consistently at the limit of what the series can do, Sol is the one to test against.

How you reach it

The API, the apps it powers, and what its limits let you do.

Terra is reached through the OpenAI API and serves as a general workhorse behind coding tools, reasoning tasks and agent loops. The context window runs to just over a million tokens, which is enough to hold an entire codebase, a long document set or a lengthy agent trace in a single call, and the output ceiling is generous enough for long files and full reports rather than fragments. File and image input mean a request can carry the source material with it.

Why it matters

What changes because this exists, or why it does not.

The practical change is that the awkward middle of a workload no longer needs routing. Teams that previously split traffic between a cheap model and a flagship, with logic to decide which got what, can put the varied majority on Terra and keep the other two tiers for the genuine extremes. For long-context agentic runs, the combination of a very large window and a high output ceiling removes a chunking step that used to be its own engineering problem.

Follows GPT-5.6 Sol Pro. Superseded by GPT-5.6 Terra Pro. See the whole line.

03 / evidence

How much of this is verified

Split by category, so a strong number never hides a thin evidence base. Verified means we read it on the benchmark's own published results; a provider's claim about its own model is shown and labelled rather than dropped.

No published benchmark scores for this model yet.

We publish a score only where we can link the result to where it was published. Until a benchmark result for this model exists in a source we read, this section stays empty rather than being filled with a provider's marketing figure.

How we decide what counts as evidence

04 / ledger

Benchmark ledger

Every published row, grouped by category, each compared with the best published score on the same benchmark. 'Is 64% good' is a question nobody can answer; '26 points behind the leader' is one anybody can.

Nothing in the ledger yet.

Each row here carries a score, the benchmark it came from, what the leading model scored on the same test, and a link to the published result. Rows appear as results are published and read.

How we decide what counts as evidence

05 / capability

Capability shape

Where this model is strong, and against how many peers. Ranks are against models with evidence in that category, not against everything we track: ranking against models nobody tested would rank who published, not who is better.

No category scores to shape yet.

A category score is the weighted mean of the benchmarks published for it. With no published rows there is nothing to average, and an empty chart drawn at zero would say something false.

How we decide what counts as evidence

06 / cost

What it costs

List API rates as last read from the provider, with the source on every row, plus every change we have recorded since we started tracking it.

GPT-5.6 Terra API pricingSurge45°
ChargePriceUnitRead onSource
Input$2.00per 1M tokens2026-09-24Check
Output$12.00per 1M tokens2026-09-24Check

Terra costs more per call than the Luna tier below it and less than Sol above it, which is exactly what the middle of a three-tier range is for. Output is priced well above input, as is standard, so the cost of a workload depends far more on how much the model writes than on how much you feed it, and the very large context window is cheaper to exploit than it first looks.

What a month costsSurge45°
WorkloadInput / monthOutput / monthCost
A small product team20M tokens5M tokens$100.00
A busy support assistant200M tokens40M tokens$880.00
A document pipeline1000M tokens100M tokens$3,200.00

List API rates, no caching and no batch discount, which both providers offer and which change the answer a great deal. Treat these as the ceiling, not the bill.

07 / specs

Specifications

As listed by the provider's own catalogue and re-read every few hours. Anything absent is absent there too.

SpecificationSurge45°
Context window1,050,000 tokens
Maximum output128,000 tokens
Modalitiesfile, image, text
Released09/07/2026
StatusCurrent
Catalogue identifieropenai/gpt-5.6-terra

08 / lineage

Lineage

What this model replaced, what replaced it, and what else its provider has in the field.

Also from OpenAI

09 / line

The line

Every model in this family in release order, so a page from eight months ago says in one glance that two newer ones exist.

Came before

← GPT-5.6 Sol Pro

Came after

GPT-5.6 Terra Pro →
  1. 01GPT-4.114/04/2025
  2. 02GPT-4.1 Mini14/04/2025
  3. 03GPT-4.1 Nano14/04/2025
  4. 04GPT-507/08/2025
  5. 05GPT-5 Mini07/08/2025
  6. 06GPT-5 Nano07/08/2025
  7. 07GPT-5 Pro06/10/2025
  8. 08GPT-5.113/11/2025
  9. 09GPT-5.1-Codex13/11/2025
  10. 10GPT-5.1-Codex-Max04/12/2025
  11. 11GPT-5.210/12/2025
  12. 12GPT-5.2 Pro10/12/2025
  13. 13GPT-5.6 Luna09/07/2026
  14. 14GPT-5.6 Luna Pro09/07/2026
  15. 15GPT-5.6 Sol09/07/2026
  16. 16GPT-5.6 Sol Pro09/07/2026
  17. 17GPT-5.6 Terra09/07/2026
  18. 18GPT-5.6 Terra Pro09/07/2026
  19. 19GPT-6 Astra04/09/2026
  20. 20GPT-6 Astra Pro04/09/2026
  21. 21GPT-6 Luna22/09/2026
  22. 22GPT-6 Luna Pro22/09/2026
  23. 23GPT-6 Sol22/09/2026
  24. 24GPT-6 Sol Pro22/09/2026

Ordered by release date and worked out from the naming, so a new member slots in as soon as its page exists. A retired model keeps its page and its place in the line.

10 / notes

Our notes

What this model changes for a brand trying to be cited in AI answers, and every change we have logged since it launched.

What it changes for you

A model that can hold a million tokens of context will read your whole documentation set, not just the page that ranked. That rewards brands whose product information is complete, consistent and machine-readable across the whole site, and punishes those whose story only holds together on the landing page. File and image input also mean PDFs, spec sheets and screenshots are now part of what gets read, so they need to say the same thing your web copy does.

Where buyers meet this model

Most buyers meet Terra without being told its name, as the default model behind a coding assistant, an internal agent or a document tool their vendor built on the OpenAI API. Developers meet it directly through the API when choosing a tier within the GPT-5.6 series. Anywhere an AI answer is being assembled from retrieved sources, a model of this class is doing the reading and the summarising.

Change log

Nothing published here yet. Changes appear within hours of a provider announcing them.

11 / questions

Questions

The things people ask about this model, answered from what is on this page rather than from anywhere else.

When should I use GPT-5.6 Terra instead of Sol or Luna?
Use Terra when your workload is mixed and you cannot predict in advance which requests will be demanding. Luna is the better choice for high-volume simple tasks, and Sol for work that consistently sits at the hardest end of what the series can handle.
What can GPT-5.6 Terra actually take as input?
Text, images and files. That means a request can include a document or a screenshot directly, rather than requiring a separate extraction step before the model sees it.
Does the large context window make it expensive to run?
Not especially. Input is priced considerably lower than output, so filling the context window costs less than generating long responses, and total spend tracks how much the model writes more than how much you send it.
Surge45°

Is GPT-5.6 Terra recommending you?

Models change what gets cited. We measure whether AI answers name your brand or your competitors across every assistant, and show you what to change.

Free AI visibility audit Book a call