stablelm-2-1.6B-chat: benchmark results
Stability AI's StableLM 2 Chat 1.6B, an instruction-tuned fine-tune of StableLM 2 1.6B using DPO in a pipeline inspired by Hugging Face's Zephyr 7B, with ChatML-formatted conversations. Provider: Stability AI. Released 2024-04-08. Access: Open.
Aggregate Rating 1278 ± 32, rank #1487 of 1635 rated models, from 17 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - GSM8K | 38.82 | Accuracy (%) (5-shot) | 53.6 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 46.5 | MC2 (%) (0-shot) | 34.7 |
| Open Portuguese LLM - ENEM | 40.31 | Accuracy (%) | 29.8 |
| Open Portuguese LLM - OAB Exams | 32.85 | Accuracy (%) | 29.2 |
| Open Portuguese LLM - BLUEX | 33.24 | Accuracy (%) | 28.8 |
| Open LLM Leaderboard - IFEval | 30.6 | Score | 28.1 |
| Open Portuguese LLM - FaQuAD NLI | 44.66 | Macro F1 (%) | 25.8 |
| Open Portuguese LLM - ASSIN2 RTE | 69.58 | Macro F1 (%) | 24.8 |
| Open LLM Leaderboard v1 - MMLU | 41.47 | Accuracy (%) (5-shot) | 21.6 |
| Open LLM Leaderboard v1 - HellaSwag | 69.24 | Normalized accuracy (%) (10-shot) | 19.9 |
| Open LLM Leaderboard v1 - ARC Challenge | 43.52 | Normalized accuracy (%) (25-shot) | 19.3 |
| Open LLM Leaderboard v1 - WinoGrande | 64.72 | Accuracy (%) (5-shot) | 17.7 |
Interactive version: theaggregate.ai/model?slug=stablelm-2-1-6b-chat · How It Works · Data refreshed daily, snapshot 2026-11-05.