GPT Terra is OpenAI’s model for clear rewrites, scheduling logic and technical explanations. Sign up to use GPT Terra side by side with 100+ other AI models.
GPT Terra is good at plain-English rewrites, constraint puzzles and short technical explanations. In our tests, GPT 5.6 Terra rewrote a legal cancellation clause in one sentence, built a valid interview schedule and checked each constraint, and explained when to store UTC versus local times.
Rewrite this policy in plain English without changing its meaning: Requests received subsequent to the published cancellation deadline shall not qualify for reimbursement except where required by applicable law.
Original
Plain English
Cancellation requests received after the published deadline are not eligible for a refund, unless applicable law requires one.
Schedule three 30-minute interviews between 9:00 and 11:00. Alex cannot start before 9:30, Bo cannot start after 10:00, and Casey is available throughout. Give a valid schedule and check each constraint.
Compare storing timestamps as UTC with storing local wall-clock times for recurring appointments. Give one case where each representation is necessary.
Make writing clearer
Answered November 4, 2026, shown unedited
Rewrite this policy in plain English without changing its meaning: Requests received subsequent to the published cancellation deadline shall not qualify for reimbursement except where required by applicable law. Keep your answer under 180 words, with code where useful.
GPT 5.6 Terra
Cancellation requests received after the published deadline are not eligible for a refund, unless applicable law requires one.
Solve a scheduling problem
Answered November 4, 2026, shown unedited
Schedule three 30-minute interviews between 9:00 and 11:00. Alex cannot start before 9:30, Bo cannot start after 10:00, and Casey is available throughout. Give a valid schedule and check each constraint. Keep your answer under 180 words, with code where useful.
GPT 5.6 Terra
| Time | Candidate |
|---|---|
| 9:00–9:30 | Bo |
| 9:30–10:00 | Alex |
| 10:00–10:30 | Casey |
Constraint check:
Alex: starts 9:30 ✅ not before 9:30
Bo: starts 9:00 ✅ not after 10:00
Casey: available throughout ✅
All interviews: 30 minutes each, within 9:00–11:00 ✅
Explain a technical choice
Answered November 4, 2026, shown unedited
Compare storing timestamps as UTC with storing local wall-clock times for recurring appointments. Give one case where each representation is necessary. Keep your answer under 180 words, with code where useful.
GPT 5.6 Terra
UTC timestamps represent a specific instant globally.
2025-03-08T15:00:00Z
Use UTC when the event must happen at the same moment everywhere—for example, a webinar, system job, payment deadline, or audit log. A webinar at 15:00Z appears as different local times but starts simultaneously for all attendees.
Local wall-clock times represent a clock reading in a particular time zone.
Every Monday at 09:00
Zone: America/New_York
Use this for recurring human schedules that should remain at the same local time—for example, a weekly doctor appointment, school opening time, or store hours. Around daylight-saving changes, its UTC equivalent changes, but it remains 9:00 AM locally.
For recurring appointments, store:
{
"weekday": "MO",
"time": "09:00",
"timeZone": "America/New_York"
}
Then calculate each occurrence’s UTC instant using the zone’s rules. Avoid storing only a fixed UTC time for local recurrences, or DST may shift the displayed local hour.
Terra’s answers were brief, so test it on longer tasks before relying on it for depth. Three sample prompts tested live by Chandan Kumar on November 4, 2026. Answers are unedited and regenerated for every new model.
GPT-6.1 Terra costs the least, from 8 Sparks per message. GPT 5.6 Terra costs the most, from 20.
| Model | Speed | Reasoning | Context | Sparks | Plan | Est. messages a month on Pro |
|---|---|---|---|---|---|---|
| GPT-6.1 Terra Latest | Fast | Advanced | 128K | Sparks8+ | Pro | ≈1,250 |
| GPT 5.6 Terra | Standard | Advanced | 1.1M | Sparks20+ | Pro | ≈500 |
Sparks are per message and start at the amount shown. Monthly counts are estimates; longer chats and attached files use more.
Context window
The largest on Sparkian is 1.1M tokens. Attached files count toward it.
Sparks per message
Included with the Pro plan.
Est. messages a month on Pro
$19 a month with 10,000 Sparks included.
GPT-6.1 Terra is the newest GPT Terra model, released October 26, 2026. You can use both versions on Sparkian.
Compare GPT Terra with Claude Fable and Claude Opus
See how GPT Terra holds up against other models in one prompt.
One plan, 100+ models
No separate subscription per provider.
Research with sources
Web search and reading, cited.
Images and teams
Generate in chat, share projects.
100+ models from more than 20 AI providers, all on one plan.
Pricing, plans and which GPT Terra model to pick.
Compare GPT Terra with these in one chat.
GPT is a trademark of OpenAI. Sparkian is not affiliated with or endorsed by model providers.
Your next question, first draft, or big idea. Start here.
