DeciLM-7B-instruct: benchmark results
Deci AI's DeciLM-7B-instruct, a LoRA fine-tune of DeciLM-7B on the SlimOrca dataset for short-form instruction following, without RLHF or DPO preference optimization. Provider: Deci AI. Released 2023-12-10. Access: Open.
Aggregate Rating 1395 ± 26, rank #1239 of 1656 rated models, from 18 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - WinoGrande | 79.72 | Accuracy (%) (5-shot) | 70.7 |
| Open LLM Leaderboard v1 - GSM8K | 46.02 | Accuracy (%) (5-shot) | 59.8 |
| Open CoT - LSAT Reading Comprehension | 13.75 | CoT Gain (%) | 57.6 |
| Open LLM Leaderboard - IFEval | 48.8 | Score | 57 |
| Open CoT - LSAT Logical Reasoning | 10.78 | CoT Gain (%) | 53.1 |
| Open LLM Leaderboard v1 - ARC Challenge | 61.01 | Normalized accuracy (%) (25-shot) | 50.8 |
| Open LLM Leaderboard v1 - HellaSwag | 82.37 | Normalized accuracy (%) (10-shot) | 50.8 |
| Open CoT - LogiQA 2 | 6.62 | CoT Gain (%) | 49.6 |
| Open CoT Leaderboard | 7.17 | Average CoT Gain (%) | 48.1 |
| Open LLM Leaderboard v1 - MMLU | 60.24 | Accuracy (%) (5-shot) | 47.3 |
| Open CoT - LogiQA | 3.83 | CoT Gain (%) | 46.2 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 49.75 | MC2 (%) (0-shot) | 45.9 |
Interactive version: theaggregate.ai/model?slug=decilm-7b-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.