smol_llama-101M-GQA [bfloat16] [rev: cac68b3377fd]: benchmark results
BEE-spoke-data release at revision cac68b3377fd0a1eb1aca92a2e661d81f59d8b08: a six-layer Llama-architecture decoder with 24 query heads and eight key/value heads. This catalog row retains its explicit bfloat16 evaluation mode; configured storage dtype does not establish evaluator dtype. Provider: BEE-spoke-data. Access: Open.
Aggregate Rating 1098 ± 39, rank #10736 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 45.76 | MC2 (%) (0-shot) | 32.1 |
| Open LLM Leaderboard v1 - GSM8K | 0.83 | Accuracy (%) (5-shot) | 14.3 |
| Open LLM Leaderboard v1 - HellaSwag | 28.77 | Normalized accuracy (%) (10-shot) | 4.9 |
| Open LLM Leaderboard v1 - WinoGrande | 50.67 | Accuracy (%) (5-shot) | 4.1 |
| Open LLM Leaderboard v1 - MMLU | 24.24 | Accuracy (%) (5-shot) | 3.3 |
| Open LLM Leaderboard v1 - ARC Challenge | 23.55 | Normalized accuracy (%) (25-shot) | 2.8 |
Interactive version: theaggregate.ai/model?slug=smol-llama-101m-gqa-bfloat16-rev-cac68b3377fd · How It Works · Data refreshed daily, snapshot 2026-11-08.