Decision model

REAX 1.3

1 question per request

Accuracy, all 2,198 questions
77.9%76.2% to 79.6%, 95% interval; 1,713 right
Answers kept at 99% confidence
240 of 2,1980 wrong; share right 98.4% to 100.0%, 95% interval
Cost, $ per 1,000 questionsper 1M questions
$0.185$185Basis and price under Cost
Latency
0.250 smedian request, 1 question per request

Back to the leaderboard.

How we called it

Kind
Decision model
Access
REAX Engine server (reax13 mode, the maker's release code), self-hosted, FP8
Paid as
Our own hardware
Model ID
reax13
Parameters
Two models run together, option probabilities averaged 50/50: 57.36B in all, all used on every question.two full models: part G (Gemma 4 31B line, with vision encoder) + part Q (Qwen3.8-27B line, with vision encoder and 255-way readout); source: counted from the checkpoint's safetensors, decisionmodels/REAX-1.3@fa4bf7e, huggingface.co/decisionmodels/REAX-1.3; read 9 November 2026
Hardware
1 × RTX PRO 6000 Blackwell (96 GB); FP8 (block-128)
Note
REAX's released REAX 1.3 weights (decisionmodels/REAX-1.3 at fa4bf7e), a decision model of two parts that both score every question; each option's probability is the 50/50 mean of the two parts'. Part G blends CharlieBT/gx31-decider (Gemma line, from tattrongvu/Jeffan-31b-1710) with PelaAI/KnowLine-31B-Gen1 (Gemma-4-31B); part Q blends gurujustin/sn92-gen-p5 (Qwen3.5-27B line, through baselabs/mammoth-1) with vllm-sr/d3-ultra. The model card calls REAX a decentralised lab on Bittensor subnet 92. It ran on our hardware through the maker's engine (the reax13 mode of reaxlabs/reax-engine at ff8f595, its own server, serve.py), both parts in FP8 (block-128) on one RTX PRO 6000 Blackwell, text only, one request at a time. Its card lists English only.
Confidence from
Option score
Run
827, 6 November 2026

Settings

Reasoning
off
Request format
1 question per request
Requests at once
1
Temperature
not recorded
Run record: harness seed, harness commit and prompt version
Harness seed
0
Harness commit
d34eca45758f
Prompt version
254597b8be16

Results

QuestionsAccuracy95% intervalRight
All questions77.9%76.2% to 79.6%1,713 of 2,198
Reading92.4%90.1% to 94.2%606 of 656
Grammar71.8%69.5% to 74.0%1,107 of 1,542

Comparison group: rt. Compare with another setup.

Accuracy by topic
TopicPartAccuracyRight
Word and phrase meaningsReading89.3%117 of 131
Sentence meaningsReading94.8%146 of 154
How a writer explains and develops ideasReading87.1%61 of 70
Topic and main point of a paragraphReading95.9%94 of 98
How a paragraph fits togetherReading90.6%96 of 106
Details that support a paragraph’s main pointReading94.8%92 of 97
Word types, such as nouns and verbsGrammar75.0%105 of 140
Groups of words built around a nounGrammar73.8%48 of 65
Verbs and their formsGrammar75.8%69 of 91
Verb forms used like nouns or modifiersGrammar86.1%62 of 72
Word endings and what they doGrammar75.9%60 of 79
How words are builtGrammar56.5%39 of 69
The roles of words and phrases in a sentenceGrammar71.0%103 of 145
How verbs show who acts and who receives the actionGrammar82.4%61 of 74
Types of sentencesGrammar83.5%71 of 85
Questions combining several grammar topicsGrammar60.5%133 of 220
Sound patterns and changesGrammar39.0%30 of 77
SpellingGrammar72.9%97 of 133
PunctuationGrammar75.2%88 of 117
Problems in sentence wordingGrammar80.6%141 of 175

English

The Text tab's 547 kept reading questions translated into English, asked with this setup's Text tab settings. Every setup: English reading; against Turkish: English and Turkish.

Right answers
506 of 547, first English run (829)
Accuracy
92.5%
Non-hallucination with “I don't know”
6.7%: 3 “I don't know” of 45 not right (run 831)
Answered with “I don't know”
99.5%: 544 of 547
Error rate with “I don't know”
7.7%: 42 of 544 answers
Cost, $ per 1,000 questionsper 1M questions, first English run
0.169169from GPU time
Latency, median request, first English run
0.229 s; 1 request at once
English run
829
Machine, English runs
RTX PRO 6000
Date
6 November 2026
Request format
1 question per request
Reasoning
off
Prompt version
47f5776e71db
Translations, SHA-256
bf0d47cd385a93050c65906977e199ef7e45d5848dbbdef2bc8d124f4d5abf68

Confidence

Answers kept at 99% or more
240, of which 0 wrong (share right: 98.4% to 100.0%, 95% interval); 1,958 skipped, the model less than 99% sure
Calibration gap
4.1 points
AUROC, all questions
0.87 0.85 to 0.88
AUROC, reading
0.94 0.91 to 0.96
AUROC, grammar
0.83 0.81 to 0.85
Brier score
0.121
Accuracy of its surest answers
most confident fifth: 100.0% right; most confident half: 97.5% right; most confident four fifths: 87.1% right
Typical score of a right answer
90.7%
Typical score of a wrong answer
48.4%

Repeats

We did not repeat this setup.

Latency

Latency, median request
0.250 s (this run)
Latency, 90th percentile
0.361 s
Requests at once
1
Throughput
224.8 questions per minute of request time
Request time
9.8 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

Cost per 1,000 questionsper 1M questions
$0.185$185
Whole run
$0.41
Basis
Estimated from GPU time: the run's request time x the GPUs its server held x Runpod's on-demand price for an RTX PRO 6000, $2.49 per GPU-hour (read 5 November 2026). A cloud-equivalent estimate, not measured spending.
Price
$2.49 per GPU-hour, 1 GPU, for the run's request time (source, 2026-11-05)
Input tokens per question
0
Output tokens per question
0

Going first in front of an LLM

Replay of recorded answers, cutoff chosen on half the questions, averaged over the splits. Savings are estimates from API prices; the Method page explains how.

LLM behind itShare takenAccuracy changeSaving
Gemini 3.8 Flash high · chat33.6%−0.09 points26%
GPT-6 Astra high · chat36.4%−0.14 points33%
GPT-6.1 Sol high · chat38.2%−0.17 points27%
GPT-6 Astra low · chat37.9%−0.18 points34%
GPT-6 Astra low · chat · 1 per request34.4%−0.13 points34%
Gemini 3.8 Flash low · chat · 1 per request35.5%−0.10 points16%
GPT-6.1 Sol low · chat · 1 per request36.5%−0.14 points33%
GPT-6.1 Sol low · chat37.9%−0.18 points22%
Claude Opus 5.5 high · chat41.9%−0.22 points39%
GPT-6.1 Terra high · chat41.6%−0.20 points6%
Claude Opus 5.5 low · chat · 1 per request41.2%−0.24 points40%
Gemini 3.8 Flash low · chat41.4%−0.20 points-27%
GPT-6.1 Terra low · chat41.7%−0.21 points-2%
Claude Opus 5.5 low · chat50.5%−0.20 points47%
Claude Sonnet 5.5 high · chat · 1 per request56.0%−0.27 points54%
Claude Sonnet 5.5 high · chat76.8%−0.05 points72%