What are GPT-6.1 Terra's benchmark scores?

As of 27 Oct 2026, GPT-6.1 Terra (max effort) scores 65.9 ±7.0 on the BenchLeader Index, top 50, strongest in reasoning (86) and weakest in long context (66). That rests on 5 independent benchmark results, and every score on this page links to its publisher.

How does GPT-6.1 Terra rank?

GPT-6.1 Terra (max effort) is #36 of 790 ranked configurations on 27 Oct 2026 (top 50). 150 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.4.

How much does GPT-6.1 Terra cost per 1M tokens?

As of 27 Oct 2026, GPT-6.1 Terra costs $1.00 per 1M input tokens and $5.00 per 1M output tokens, $2.00 blended (three input tokens per output token).

Is GPT-6.1 Terra better than GPT-5.5?

As of 27 Oct 2026, GPT-5.5 (xhigh) and GPT-6.1 Terra (max) are neck and neck on quality (66.0 vs 65.9 on the BenchLeader Index, inside the ±7.9 margin of error), and GPT-6.1 Terra (max) costs 5.6× less, which makes it the better buy for most workloads. The head-to-head page has the benchmark-by-benchmark detail.

When was GPT-6.1 Terra released?

OpenAI released GPT-6.1 Terra on 26 Oct 2026. BenchLeader has results for it at 7 reasoning-effort settings.

OpenAIReasoning modelNewReleased 26 Oct 2026Compare with another model

GPT-6.1 Terra

Reasoning effort

As of 27 Oct 2026, GPT-6.1 Terra by OpenAI scores 65.9 on the BenchLeader Index (top 50 of 790 ranked configurations), with its best results in reasoning (86) and weakest in long context (66). It costs $2.00 per 1M tokens blended ($1.00 in / $5.00 out), and streams about 91 tokens/s (Artificial Analysis). Released 26 Oct 2026; 5 independent results so far.

Key facts (as of 27 Oct 2026)

GPT-6.1 Terra costs $2.00 per 1M tokens blended ($1.00 in / $5.00 out) (list price, 27 Oct 2026).
GPT-6.1 Terra streams about 91 tokens/s (Artificial Analysis, 27 Oct 2026).
GPT-6.1 Terra was released on 26 Oct 2026 (OpenAI).

In detail. GPT-6.1 Terra is an OpenAI proprietary reasoning model, released 26 Oct 2026. Its best configuration (max reasoning effort) ranks #36 of 790 on the BenchLeader Index at 65.9 ±7.0, in the upper half. The ± is the point: 150 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in reasoning (86) and lowest in long context (66). At $2.00 per million tokens blended it is pricier than most ranked models. Output speed of 91 tokens per second puts it faster than most, with a first answer in 131.8 s. It has been measured at 7 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 26 Oct 2026.

Blended price: $2.00/M; $1.00 in · $5.00 out
Output speed: 91 tok/s; measured by Artificial Analysis
First answer: 132 s; first token 2.03 s
Context: 1.1M
Full answer: 137 s; median, reasoning included
Cost per run: $0.403; one full Intelligence Index run
Released: 26 Oct 2026

How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.

Overall index66
Reasoning86
Knowledge75
Multimodal68
Long context66
Composite83

Versions

OpenAI has shipped 2 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.

Model	Released	Index	Rank
NewGPT-6.1 Terramaxthis page	26 Oct 2026	65.9	#36
GPT-5.6 Terramax	9 Jul 2026	64.2	#54

Reasoning-effort configurations

The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.

Effort	Index	Rank	Speed	First answer	Chat reply cost	Categories
no reasoning	55.6	#228	91 tok/s	1.12 s	$0.0019	Composite 67 · Knowledge 63 · Long context 59 · Multimodal 59 · Reasoning 49
low	60.9	#107	91 tok/s	1.29 s	$0.0019	Composite 70 · Knowledge 72 · Long context 65 · Multimodal 65 · Reasoning 65
medium	63.1	#74	91 tok/s	1.43 s	$0.0019	Composite 75 · Knowledge 73 · Long context 65 · Multimodal 66 · Reasoning 76
high	64.7	#46	91 tok/s	6.78 s	$0.0019	Composite 78 · Knowledge 74 · Long context 66 · Multimodal 67 · Reasoning 83
xhigh	65.0	#42	91 tok/s	18 s	$0.0019	Composite 80 · Knowledge 74 · Long context 66 · Multimodal 67 · Reasoning 84
maxbest	65.9	#36	91 tok/s	132 s	$0.0019	Composite 83 · Knowledge 75 · Long context 66 · Multimodal 68 · Reasoning 86
not stated	–	–	48 tok/s	2.03 s	$0.0019	Composite 63

What more thinking costs

Turning the effort up buys index points and multiplies the bill. Here max costs 7.4× low for +5.0 on the index — 1.7 points per doubling of spend. Compare that with other models.

Benchmark results

See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

Benchmark	no reasoning	low	medium	high	xhigh	max	not stated	Source
LiveBench 2.0 Reasoningnot in indexeffort not stated	–	–	–	–	–	–	88.7%#15	LiveBench
Humanity's Last Exam (AA)not in index	21.1%#215	35.8%#118	40.2%#89	43.3%#63	43.7%#57	46.3%#46	–	Artificial Analysis
CritPt	4.9%#163	15.1%#97	21.4%#61	25.7%#43	26.6%#38	28.0%#28	–	Artificial Analysis

Coding

Benchmark	no reasoning	low	medium	high	xhigh	max	not stated	Source
LiveBench 2.0 Codingnot in indexeffort not stated	–	–	–	–	–	–	90.8%#8	LiveBench
SciCode (AA)not in index	47.8%#134	52.4%#88	51.6%#96	51.6%#96	53.4%#78	55.1%#54	–	Artificial Analysis
Terminal-Bench 4.0 (AA)not in index	26.8%#53	15.2%#75	32.3%#47	41.9%#31	46.5%#24	55.6%#12	–	Artificial Analysis
FrontierCode 1.1 (AA)not in index	41.0%#32	36.3%#47	43.7%#19	37.2%#43	41.4%#31	36.0%#48	–	Artificial Analysis

Agents & tools

Benchmark	no reasoning	low	medium	high	xhigh	max	not stated	Source
LiveBench Long Agenticnot in indexeffort not stated	–	–	–	–	–	–	48.6%#8	LiveBench
LiveBench Browsernot in indexeffort not stated	–	–	–	–	–	–	93.1%#2	LiveBench
AutomationBenchnot in index	51.0%#82	56.3%#59	57.8%#51	58.5%#46	59.9%#38	60.2%#35	–	Artificial Analysis
GDP.pdf (AA)not in index	23.2%#49	25.6%#32	28.4%#15	28.0%#16	26.6%#24	28.6%#14	–	Artificial Analysis
AA-Briefcase v1.1not in index	–	–	–	–	–	1402#29	–	Artificial Analysis
Terminal-Bench-Science 0.1 (AA)not in index	24.3%#24	7.6%#48	24.8%#23	33.3%#19	41.9%#14	57.6%#5	–	Artificial Analysis

Knowledge

Benchmark	no reasoning	low	medium	high	xhigh	max	not stated	Source
LiveBench 2.0 Data Analysisnot in indexeffort not stated	–	–	–	–	–	–	82.3%#4	LiveBench
LiveBench Researchnot in indexeffort not stated	–	–	–	–	–	–	73.9%#17	LiveBench
AA-Omniscience	3.9#120	23.7#59	27.1#48	28.4#42	28.7#39	29.8#33	–	Artificial Analysis
AA-Omniscience: accuracynot in index	44.0%#101	47.8%#79	49.4%#69	50.0%#62	49.6%#66	49.1%#70	–	Artificial Analysis
AA-Omniscience: non-hallucinationnot in index	28.2%#233	54.0%#107	56.0%#94	56.9%#90	58.5%#85	62.0%#72	–	Artificial Analysis

Multimodal

Benchmark	no reasoning	low	medium	high	xhigh	max	not stated	Source
LiveBench Designnot in indexeffort not stated	–	–	–	–	–	–	82.3%#7	LiveBench
MMMU-Pro	76.0%#99	82.5%#37	82.8%#33	83.8%#27	83.8%#27	84.7%#17	–	Artificial Analysis

Long context

Benchmark	no reasoning	low	medium	high	xhigh	max	Source
AA-LCR	67.7%#244	81.0%#64	80.7%#69	82.3%#42	81.7%#54	82.7%#36	Artificial Analysis

Composite

Benchmark	no reasoning	low	medium	high	xhigh	max	not stated	Source
LiveBench 2.0effort not stated	–	–	–	–	–	–	77.2%#7	LiveBench
AA Intelligence Index v4.3.2	33.5#112	36.2#90	41.3#62	43.9#48	45.4#38	48.1#26	–	Artificial Analysis

Safety & honesty

Benchmark	not stated	Source
FORTRESSnot in index	18.2%#43	Scale AI SEAL

Where to run it

Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.

Provider	Speed	First token	Input $/M	Output $/M	Context	Quantisation
OpenAI	65 tok/s	2.20 s	$0.500	$2.50	1.1M	–
OpenAI	51 tok/s	1.40 s	$2.00	$10.00	1.1M	–
OpenAI	48 tok/s	2.03 s	$1.00	$5.00	1.1M	–

What a task costs

Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.050 per 1M. Reasoning tokens are not modelled.

Workload	Tokens in / out	Cost	With caching	Time
Chat reply	400 / 300	$0.0019	$0.0016	2.3 min
Summarize a 30-page report	12,000 / 600	$0.015	$0.0065	2.3 min
Code edit	6,000 / 1,500	$0.013	$0.0092	2.5 min
Agentic coding session	60,000 / 4,000	$0.080	$0.037	2.9 min
Structured extraction	2,000 / 200	$0.0030	$0.0016	2.2 min

New this month

Compare GPT-6.1 Terra with…

Questions people ask

What are GPT-6.1 Terra's benchmark scores?: As of 27 Oct 2026, GPT-6.1 Terra (max effort) scores 65.9 ±7.0 on the BenchLeader Index, top 50, strongest in reasoning (86) and weakest in long context (66). That rests on 5 independent benchmark results, and every score on this page links to its publisher.
How does GPT-6.1 Terra rank?: GPT-6.1 Terra (max effort) is #36 of 790 ranked configurations on 27 Oct 2026 (top 50). 150 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.4.
How much does GPT-6.1 Terra cost per 1M tokens?: As of 27 Oct 2026, GPT-6.1 Terra costs $1.00 per 1M input tokens and $5.00 per 1M output tokens, $2.00 blended (three input tokens per output token).
Is GPT-6.1 Terra better than GPT-5.5?: As of 27 Oct 2026, GPT-5.5 (xhigh) and GPT-6.1 Terra (max) are neck and neck on quality (66.0 vs 65.9 on the BenchLeader Index, inside the ±7.9 margin of error), and GPT-6.1 Terra (max) costs 5.6× less, which makes it the better buy for most workloads. The head-to-head page has the benchmark-by-benchmark detail.
When was GPT-6.1 Terra released?: OpenAI released GPT-6.1 Terra on 26 Oct 2026. BenchLeader has results for it at 7 reasoning-effort settings.

Data as of 27 Oct 2026. Compare these configurations.

Cite as: BenchLeader, “GPT-6.1 Terra: benchmarks, pricing, speed and rank”, data as of 27 Oct 2026, https://www.benchleader.com/models/gpt-6-1-terra.