Chat model

Kimi K3.1 Pro

reasoning none · 16 questions per request

72.7% of the 2,198 questions right (70.8% to 74.5%, 95% interval). Back to the leaderboard.

How we called it

Kind: Chat model
Access: OpenRouter chat API, Moonshot AI's endpoint (mxfp4), fallbacks off
Paid as: Provider service
Model ID: moonshotai/kimi-k3.1-pro
Confidence from: Written estimate
Run: 315, 28 October 2026

Settings

Reasoning: none
Request format: 16 questions per request
Requests at once: 4
Temperature: not recorded
Harness seed: 0
Harness commit: 001ce5b42634
Prompt version: 254597b8be16

Results

Questions	Accuracy	95% interval	Right
All questions	72.7%	70.8% to 74.5%	1,598 of 2,198
Reading	92.5%	90.3% to 94.3%	607 of 656
Grammar	64.3%	61.8% to 66.6%	991 of 1,542

Comparison group: o. Compare with another setup.

Accuracy by topic

Topic	Part	Accuracy	Right
Word and phrase meanings	Reading	90.8%	119 of 131
Sentence meanings	Reading	96.1%	148 of 154
How a writer explains and develops ideas	Reading	82.9%	58 of 70
Topic and main point of a paragraph	Reading	98.0%	96 of 98
How a paragraph fits together	Reading	85.8%	91 of 106
Details that support a paragraph’s main point	Reading	97.9%	95 of 97
Word types, such as nouns and verbs	Grammar	67.1%	94 of 140
Groups of words built around a noun	Grammar	55.4%	36 of 65
Verbs and their forms	Grammar	75.8%	69 of 91
Verb forms used like nouns or modifiers	Grammar	70.8%	51 of 72
Word endings and what they do	Grammar	64.6%	51 of 79
How words are built	Grammar	50.7%	35 of 69
The roles of words and phrases in a sentence	Grammar	61.4%	89 of 145
How verbs show who acts and who receives the action	Grammar	82.4%	61 of 74
Types of sentences	Grammar	68.2%	58 of 85
Questions combining several grammar topics	Grammar	50.5%	111 of 220
Sound patterns and changes	Grammar	31.2%	24 of 77
Spelling	Grammar	72.9%	97 of 133
Punctuation	Grammar	65.0%	76 of 117
Problems in sentence wording	Grammar	79.4%	139 of 175

Confidence

Answers kept at 99% or more: 442, of which 36 wrong (share right: 88.9% to 94.1%, 95% interval); 1,756 skipped, the model less than 99% sure
Calibration gap: 20.1 points
AUROC, all questions: 0.73 0.71 to 0.75
AUROC, reading: 0.77 0.70 to 0.85
AUROC, grammar: 0.68 0.65 to 0.70
Brier score: 0.221
Accuracy of its surest answers: most confident fifth: 91.9% right; most confident half: 85.3% right; most confident four fifths: 79.6% right
Typical score of a right answer: 97.0%
Typical score of a wrong answer: 90.0%

Repeats

We did not repeat this setup.

Latency

Latency, median request: 5.75 s (this run)
Latency, 90th percentile: 7.65 s
Requests at once: 4
Throughput: 625.6 questions per minute of request time
Request time: 3.5 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: $0.85
Whole run: $1.88
Basis: Estimated from published prices and recorded token usage
Price: $2 input, $0.20 cached input, $10 output per 1M tokens (source, 2026-10-28)
Input tokens per question: 354
Output tokens per question: 14