Decision model
Celeris-1 Decision
1 question per request
60.5% of the 2,198 questions right (58.4% to 62.5%, 95% interval). Back to the leaderboard.
How we called it
- Kind
- Decision model
- Access
- Celeris System One API
- Paid as
- Provider service
- Model ID
celeris-1-decision- Confidence from
- Option score
- Run
- 150, 10 October 2026
Settings
- Reasoning
- off
- Request format
- 1 question per request
- Requests at once
- 1
- Temperature
- not recorded
- Harness seed
- 0
- Harness commit
- not recorded
- Prompt version
- not recorded
Results
| Questions | Accuracy | 95% interval | Right |
|---|---|---|---|
| All questions | 60.5% | 58.4% to 62.5% | 1,329 of 2,198 |
| Reading | 85.2% | 82.3% to 87.7% | 559 of 656 |
| Grammar | 49.9% | 47.4% to 52.4% | 770 of 1,542 |
Comparison group: uvwxy. Compare with another setup.
Accuracy by topic
Confidence
- Answers kept at 99% or more
- 0, of which 0 wrong; 2,198 skipped, the model less than 99% sure
- Calibration gap
- 4.3 points
- AUROC, all questions
- 0.81 0.79 to 0.82
- AUROC, reading
- 0.88 0.85 to 0.90
- AUROC, grammar
- 0.73 0.71 to 0.76
- Brier score
- 0.175
- Accuracy of its surest answers
- most confident fifth: 98.6% right; most confident half: 81.8% right; most confident four fifths: 68.4% right
- Typical score of a right answer
- 77.0%
- Typical score of a wrong answer
- 44.0%
Repeats
Run 5 times on all questions.
- Answers changed in 5 runs
- 206 of 2,198
- Same wrong answer in all 5
- 738 of 2,198
- Wrong in at least one run
- 944 of 2,198
- Accuracy in each run
- 60.5%, 60.3%, 60.1%, 60.0%, 59.8%
Latency
- Latency, median request
- 0.401 s (this run)
- Latency, 90th percentile
- 0.418 s
- Requests at once
- 1
- Throughput
- 149.0 questions per minute of request time
- Request time
- 14.8 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
- $ per 1,000 questions
- $0.015
- Whole run
- $0.034
- Basis
- Estimated from published prices and recorded token usage
- Price
- $0.04 input, output free per 1M tokens (source, 2026-10-10)
- Input tokens per question
- 382
- 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 it | Share taken | Accuracy change | Saving |
|---|---|---|---|
| Gemini 3.8 Flashhigh reasoning · 16 questions per request | 17.4% | −0.12 points | 17% |
| GPT-6 Astrahigh reasoning · 16 questions per request | 15.9% | −0.13 points | 16% |
| GPT-6.1 Solhigh reasoning · 16 questions per request | 15.9% | −0.13 points | 15% |
| GPT-6 Astralow reasoning · 16 questions per request | 15.9% | −0.13 points | 16% |
| GPT-6 Astralow reasoning · 1 question per request | 17.2% | −0.14 points | 17% |
| Gemini 3.8 Flashlow reasoning · 1 question per request | 14.2% | −0.11 points | 13% |
| GPT-6.1 Sollow reasoning · 1 question per request | 15.9% | −0.13 points | 16% |
| GPT-6.1 Sollow reasoning · 16 questions per request | 15.9% | −0.13 points | 15% |
| Claude Opus 5.5high reasoning · 16 questions per request | 17.7% | −0.14 points | 17% |
| GPT-6.1 Terrahigh reasoning · 16 questions per request | 19.5% | −0.15 points | 17% |
| Claude Opus 5.5low reasoning · 1 question per request | 15.7% | −0.17 points | 16% |
| Gemini 3.8 Flashlow reasoning · 16 questions per request | 20.2% | −0.12 points | 14% |
| GPT-6.1 Terralow reasoning · 16 questions per request | 21.2% | −0.12 points | 18% |
| Claude Opus 5.5low reasoning · 16 questions per request | 17.8% | −0.15 points | 17% |
| Claude Sonnet 5.5high reasoning · 1 question per request | 23.3% | −0.12 points | 23% |
| Claude Sonnet 5.5high reasoning · 16 questions per request | 33.1% | +0.01 points | 33% |