orthorus-125B-v2: benchmark results
An ibivibiv mixture-of-experts release with a published Mixtral configuration specifying two local experts, both selected per token, and 80 decoder layers. This version remains separate from the earlier Orthorus release. Provider: ibivibiv. Access: Open.
Aggregate Rating 1514 ± 34, rank #2122 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - WinoGrande | 85.48 | Accuracy (%) (5-shot) | 98.4 |
| Open LLM Leaderboard v1 - ARC Challenge | 73.63 | Normalized accuracy (%) (25-shot) | 97.9 |
| Open LLM Leaderboard v1 - MMLU | 75.99 | Accuracy (%) (5-shot) | 97.2 |
| Open LLM Leaderboard v1 - HellaSwag | 89.04 | Normalized accuracy (%) (10-shot) | 96.2 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 70.19 | MC2 (%) (0-shot) | 90.1 |
| Open LLM Leaderboard v1 - GSM8K | 68.99 | Accuracy (%) (5-shot) | 87.6 |
Interactive version: theaggregate.ai/model?slug=orthorus-125b-v2 · How It Works · Data refreshed daily, snapshot 2026-11-08.