Chat model

GPT-6.1 Terra

high reasoning · 16 questions per request

94.3% of the 2,198 questions right (93.2% to 95.2%, 95% interval). Back to the leaderboard.

Which decision model can go first in front of it? See the Routing tab.

How we called it

Kind: Chat model
Access: Codex CLI, ChatGPT subscription
Paid as: Subscription
Model ID: gpt-6.1-terra
Confidence from: Written estimate
Run: 275, 28 October 2026

Settings

Reasoning: high
Request format: 16 questions per request
Requests at once: 1
Temperature: not recorded
Harness seed: 0
Harness commit: 0fbd97b0899d
Prompt version: 254597b8be16

Reasoning effort levels (OpenAI's model page, read 28 October 2026): none, low, medium (default), high, xhigh, max. Top level: max. This run: high.

Results

Questions	Accuracy	95% interval	Right
All questions	94.3%	93.2% to 95.2%	2,072 of 2,198
Reading	96.2%	94.4% to 97.4%	631 of 656
Grammar	93.5%	92.1% to 94.6%	1,441 of 1,542

Comparison group: fg. Compare with another setup.

Accuracy by topic

Topic	Part	Accuracy	Right
Word and phrase meanings	Reading	95.4%	125 of 131
Sentence meanings	Reading	98.7%	152 of 154
How a writer explains and develops ideas	Reading	91.4%	64 of 70
Topic and main point of a paragraph	Reading	96.9%	95 of 98
How a paragraph fits together	Reading	95.3%	101 of 106
Details that support a paragraph’s main point	Reading	96.9%	94 of 97
Word types, such as nouns and verbs	Grammar	97.1%	136 of 140
Groups of words built around a noun	Grammar	95.4%	62 of 65
Verbs and their forms	Grammar	97.8%	89 of 91
Verb forms used like nouns or modifiers	Grammar	98.6%	71 of 72
Word endings and what they do	Grammar	94.9%	75 of 79
How words are built	Grammar	91.3%	63 of 69
The roles of words and phrases in a sentence	Grammar	95.2%	138 of 145
How verbs show who acts and who receives the action	Grammar	94.6%	70 of 74
Types of sentences	Grammar	97.6%	83 of 85
Questions combining several grammar topics	Grammar	88.2%	194 of 220
Sound patterns and changes	Grammar	87.0%	67 of 77
Spelling	Grammar	94.7%	126 of 133
Punctuation	Grammar	85.5%	100 of 117
Problems in sentence wording	Grammar	95.4%	167 of 175

Confidence

Answers kept at 99% or more: 1,363, of which 19 wrong (share right: 97.8% to 99.1%, 95% interval); 835 skipped, the model less than 99% sure
Calibration gap: 2.8 points
AUROC, all questions: 0.81 0.77 to 0.85
AUROC, reading: 0.93 0.87 to 0.97
AUROC, grammar: 0.78 0.73 to 0.82
Brier score: 0.049
Accuracy of its surest answers: most confident fifth: 98.8% right; most confident half: 98.6% right; most confident four fifths: 97.5% right
Typical score of a right answer: 99.0%
Typical score of a wrong answer: 94.0%

Repeats

Run 5 times on a sample of 200 questions.

Answers changed in 5 runs: 11 of 200
Same wrong answer in all 5: 5 of 200
Wrong in at least one run: 16 of 200
Accuracy in each run: 96.5%, 97.0%, 95.0%, 96.5%, 95.5%

Latency

Latency, median request: 7.10 s (this run)
Latency, 90th percentile: 12.6 s
Requests at once: 1
Throughput: 113.3 questions per minute of request time
Request time: 19.4 minutes, over the requests whose answers were kept

Latency: the time from sending one request to receiving its answer, measured from our computer, network and any retries within the request included. Request time: the time at least one request was in flight, over the requests whose answers were kept. It is not the run's elapsed time: it leaves out requests that failed and any pause before a run was resumed.

Cost

$ per 1,000 questions: $0.51
Whole run: $1.13
Basis: Subscription run priced as if we had used the API
Price: $1 input, $0.05 cached input, $1.25 cache writes, $5 output per 1M tokens (source, 2026-10-28)
Input tokens per question: 529
Output tokens per question: 30