Motif-2-12.7B: benchmark results
Motif Technologies' open 12.7B model with Grouped Differential Attention, pretrained on 5.5T tokens (November 2025). Provider: Motif. Released 2025-11-07.
Aggregate Rating 1582 ± 42, rank #1220 of 10916 rated models, from 11 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA AIME 2025 | 80.33 | Accuracy (%) | 76 |
| AA LiveCodeBench | 65.08 | Pass@1 (%) | 71.8 |
| AA IFBench | 57.01 | Accuracy (%) | 68.1 |
| AA MMLU-Pro | 79.63 | Accuracy (%) | 65.2 |
| Tau2-Bench Telecom | 46.49 | Success Rate (%) | 51.7 |
| AA TAU-2 Bench | 46.49 | Accuracy (%) | 50.2 |
| AA GPQA Diamond | 69.49 | Accuracy (%) | 47.2 |
| AA Humanity's Last Exam | 8.76 | Accuracy (%) | 45.4 |
| AA Terminal-Bench Hard | 3.79 | Accuracy (%) | 25.6 |
| AA CritPt | 0 | Accuracy (%) | 20.9 |
| AA Long Context Reasoning | 14.67 | Accuracy (%) | 17.2 |
Interactive version: theaggregate.ai/model?slug=motif-2-12-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.