SURVEY RECORD N° 06 FILED 2026-08-13
← RETURN TO THE REGISTER

GPT-5.6 Terra

The value pick inside OpenAI's lineup; undercuts the prior GPT-5.4 tier.

Context1.05M
Passage, per 1M in / out$2 / $12
AA Intelligencenot yet surveyed
Licenseclosed
ModalitiesText, image
Official siteopenai.com →

In its favor

  • Value pick of the GPT-5.6 line at $2 / $12
  • Same 1.05M-token context as the flagship
  • Cut 20% on Jul 30, 2026
  • Common production default for many teams

Against it

  • No standalone independent AA index on file
  • Below Sol on the hardest reasoning tasks
  • Closed weights — no self-hosting
  • Long-context surcharge applies at scale
balancedagenticvision1M contextclosedvalue

ASSESSMENT

Terra is the mid-tier of OpenAI's GPT-5.6 family and, for many teams, the one that actually ships to production. It shares the flagship's 1.05M-token context but lands at $2 / $12 per million tokens — a fraction of Sol's output rate — and took a further 20% cut on July 30, 2026.

There is no standalone independent intelligence score for Terra in this survey, and on the hardest reasoning tasks it sits below Sol; the long-context surcharge that applies across the family kicks in at scale here too. What Terra offers is the GPT-5.6 line's balance point: most of the capability, the full context window, and a price that makes high-volume use practical.

For teams on OpenAI's stack who don't need the absolute top of the reasoning curve, Terra is usually the default. Those who do step up to Sol; those optimizing purely for cost drop to Luna or look to the open-weight tier.

FIELD NOTE — Cut 20% on Jul 30. Production default for many teams.

VISIT OFFICIAL SITE →
⚑ POSITION NOT YET SURVEYED
LOCATOR — FIG. 1, GENERAL CHART