Apriel-v1.6-15B-Thinker: benchmark results
ServiceNow's open 15B multimodal reasoning model, an Apriel 1.5 upgrade with leaner reasoning traces. Provider: ServiceNow. Released 2025-09-01.
Aggregate Rating 1553 ± 42, rank #1663 of 10916 rated models, from 24 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA LiveCodeBench | 80.74 | Pass@1 (%) | 92.4 |
| AA AIME 2025 | 88 | Accuracy (%) | 86 |
| AA IFBench | 69.12 | Accuracy (%) | 83.3 |
| Tau2-Bench Telecom | 69.3 | Success Rate (%) | 63.7 |
| AA MMLU-Pro | 78.97 | Accuracy (%) | 62.4 |
| AA TAU-2 Bench | 69.3 | Accuracy (%) | 62.3 |
| AA GPQA Diamond | 73.33 | Accuracy (%) | 53.6 |
| AA Omniscience - Software Engineering (SWE) - JavaScript | 28.18 | Accuracy (%) | 52.9 |
| AA Terminal-Bench Hard | 14.39 | Accuracy (%) | 51.3 |
| AA Omniscience - Software Engineering (SWE) - Swift | 32 | Accuracy (%) | 51.2 |
| AA Humanity's Last Exam | 10.8 | Accuracy (%) | 50.4 |
| AA Omniscience - Software Engineering (SWE) - PHP | 20 | Accuracy (%) | 49.5 |
Interactive version: theaggregate.ai/model?slug=apriel-v1-6-15b-thinker · How It Works · Data refreshed daily, snapshot 2026-11-08.