OpenAI
GPT-6.1 Terra
OpenAI's cost-intelligence and computer-use line, released 2026-10-26 at $1/$5 per million tokens, 50% below GPT-5.6 Terra and GPT-6.1 Sol, per OpenAI, with a 1,050,000-token context and a June 1, 2026 cutoff. This site records one benchmark row: OpenAI's own Terminal-Bench 4.0, 50.3% at max effort. No independent run is recorded here as of 2026-10-28, and OpenAI's Agents' Last Exam claim is a +15pp gain at medium, not an absolute score.
GPT-6.1 Terra benchmarks and pricing, every number sourced: 1 tracked GPT-6.1 Terra benchmark score (0 independently run, 1 still resting on a vendor’s own claim), priced at $1.00 per million input tokens and $5.00 per million output.
- Released
- 2026-10-26
- Availability
- API; ChatGPT Work and Codex for Plus/Pro/Enterprise/Edu; Bedrock GA same day as launch (2026-10-26)
- License
- proprietary
- Context window
- 1M tokensvendor-stated (developers.openai.com/api/docs/models/gpt-6.1-terra: 1,050,000-token context; checked 2026-10-28)
- Knowledge cutoff
- 2026-06-01
- Verified
- source checked 2026-10-28
- Parameters
- Not disclosed
- Architecture
- Not disclosed
GPT-6.1 Terra’s verified record
GPT-6.1 Terra’s most-compared rival is Claude Fable 5.1: 1 not callable across their 1 shared comparison. GPT-6.1 Terra is announced at $1.00/$5.00 per 1M tokens (in/out) vs Claude Fable 5.1’s $10.00/$50.00.
Against the 19 head-to-head comparisons GPT-6.1 Terra shares with other tracked models:
A gap counts for GPT-6.1 Terra only where both sides were run independently and the benchmark still separates models. Losses are listed alongside wins on purpose — GPT-6.1 Terra trails on 0 of them.
No verdict for GPT-6.1 Terra anywhere on Terminal-Bench 4.0 · max ( nothing independently confirmed on both sides).
- None of GPT-6.1 Terra’s agentic comparisons are independently confirmed on both sides yet.
GPT-6.1 Terra announced pricing
$1.00 in / $5.00 out per 1M tokens — official pricing source
What GPT-6.1 Terra costs per job
| Workload | Tokens in / out | Cost |
|---|---|---|
| One long chat turn | 100K / 10K | $0.150 |
| A codebase review | 1,000K / 100K | $1.50 |
| A day of agent work | 10,000K / 1,000K | $15.00 |
Computed from GPT-6.1 Terra’s list rates above — cache discounts and batch tiers are not applied.
GPT-6.1 Terra is one of 8 OpenAI models tracked on this site, at these official list prices.
| Model | In / 1M | Out / 1M |
|---|---|---|
| GPT-6 Cyber | $20.00 | $100.00 |
| GPT-6.1 Terra | $1.00 | $5.00 |
| GPT-6.1 Sol | $2.00 | $10.00 |
| GPT-6 Luna | $0.10 | $0.50 |
| GPT-6 Sol(previous version) | $2.00 | $10.00 |
| GPT-6 Astra | $10.00 | $50.00 |
| GPT-5.6 Luna(previous version) | $0.20 | $1.20 |
| GPT-5.6 Sol(previous version) | $4.00 | $20.00 |
GPT-6.1 Terra benchmark scores
| Benchmark | Score |
|---|---|
| Terminal-Bench 4.0(max)[1] Terminal ops · ±12.4 is noise |
Who ran these numbers: 0 of 1 independent; vendor self-reported (1).
- Terminal-Bench 4.0: OpenAI launch blog, Terminal-Bench 4.0 at max reasoning effort. Two reader passes on 2026-10-28 agree on the sentence: at maximum reasoning effort, GPT-6.1 Terra solves 50.3% of tasks, against GPT-5.6 Terra's published 21.5%. Raw HTML was Cloudflare-blocked, so no embedded chart JSON was available; the number is that prose sentence. Vendor self-report, not an independent run. OpenAI's comparator, GPT-5.6 Terra at 21.5%, is not tracked here and is not stored as a row. Replace this row when vals.ai or Artificial Analysis publishes an independent Terra result; move the vendor number to notes at that point.
Notes on the record
OpenAI published GPT-6.1 Terra on 2026-10-26. Standard API pricing is $1 input / $5 output per million tokens, with cached input at $0.05 and cache write at $1.25 for 30 minutes; OpenAI says that is 50% below GPT-5.6 Terra and GPT-6.1 Sol. Prompts over 272,000 input tokens bill the whole request at 2x the input and cache rates and 1.5x output. Terra is available through the API, in ChatGPT Work and Codex for Plus/Pro/Enterprise/Edu, and on Bedrock GA the same day as launch (developers.openai.com/api/docs/models/gpt-6.1-terra, checked 2026-10-28).
Terra is a distinct OpenAI line, not a GPT-6.1 Sol tier; GPT-6.1 Sol is tracked separately. Treat the launch performance figures as vendor claims until an independent board runs the model. The tracked Terminal-Bench 4.0 row is OpenAI's 50.3% at max effort; OpenAI's comparator is 21.5% for GPT-5.6 Terra, which is not a model this site tracks and is not stored as a row. The Agents' Last Exam claim is a +15pp gain at medium versus GPT-5.6 Terra, not an absolute score, so it is not recorded as a row. OSWorld 2.1 and MMMU-Pro appear as claims in the launch blog; no numeric value for either is used here, and MMMU-Pro is not a tracked benchmark. No independent tracked Terra score exists as of 2026-10-28 (https://openai.com/index/introducing-gpt-6-1-terra, checked 2026-10-28).
Compare with
FAQ
What is GPT-6.1 Terra?
A new OpenAI model released on 2026-10-26, positioned as a distinct computer-use and cost-intelligence line rather than a tier of GPT-6.1 Sol. It accepts text and image input, returns text, and supports reasoning effort from none to max.
How much does GPT-6.1 Terra cost?
Standard API pricing is $1 per million input tokens and $5 per million output tokens. Cached input is $0.05 and cache write is $1.25 for 30 minutes. OpenAI says this is 50% below GPT-5.6 Terra and GPT-6.1 Sol. Above 272,000 input tokens the whole request bills at 2x input and cache and 1.5x output.
Has GPT-6.1 Terra been independently benchmarked?
No independent tracked row exists yet. The only recorded score is OpenAI's self-reported Terminal-Bench 4.0 result of 50.3% at max effort. OpenAI also claims a +15pp Agents' Last Exam gain at medium versus GPT-5.6 Terra, but that is a relative claim, not an absolute score, so it is not in the table.
What are GPT-6.1 Terra's context window and knowledge cutoff?
OpenAI's model documentation states a 1,050,000-token context window and a June 1, 2026 knowledge cutoff. Input is text and image, output is text, and reasoning effort ranges from none to max.
Does GPT-6.1 Terra replace GPT-6.1 Sol?
No. OpenAI presents Terra as a distinct line, and GPT-6.1 Sol remains a separate tracked model. This site does not mark either model as superseded without explicit vendor lifecycle language.
What is GPT-6.1 Terra good at?
OpenAI positions GPT-6.1 Terra for computer-use and agentic workflows, for cost-sensitive, high-volume work, and for long-context tasks. It accepts text and images, returns text, and exposes reasoning effort from none to max. This site has no independent run of those workflows yet, so treat the task-fit claims as vendor positioning rather than measured results.
Further reading
- Models with 10M token context windows 2026 — GPT-6.1 Terra is one of the 47 models it compares.
- Self-reported vs. independent AI benchmarks — the self-reported-versus-independent rule every score on GPT-6.1 Terra’s page follows.
Benchmark guides
- Terminal-Bench 4.0 — what GPT-6.1 Terra’s agentic score on it does and doesn’t prove.