GALAXY-XB-v.02: benchmark results
An experimental depth-upscaled merge released by TeeZee from SOLAR-10.7B-v1.0, removing 10 layers from each source model before merging. This revision is a merge without subsequent fine-tuning. Provider: TeeZee. Access: Open.
Aggregate Rating 1411 ± 33, rank #5569 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - MMLU | 64.99 | Accuracy (%) (5-shot) | 79.8 |
| Open LLM Leaderboard v1 - WinoGrande | 80.27 | Accuracy (%) (5-shot) | 73.2 |
| Open LLM Leaderboard v1 - HellaSwag | 83.27 | Normalized accuracy (%) (10-shot) | 57.3 |
| Open LLM Leaderboard v1 - GSM8K | 42.08 | Accuracy (%) (5-shot) | 56.8 |
| Open LLM Leaderboard v1 - ARC Challenge | 60.67 | Normalized accuracy (%) (25-shot) | 48.9 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 43.6 | MC2 (%) (0-shot) | 24.1 |
Interactive version: theaggregate.ai/model?slug=galaxy-xb-v-02 · How It Works · Data refreshed daily, snapshot 2026-11-08.