Code-290k-6.7B-Instruct: benchmark results
A full fine-tune by ajibawa-2023 of DeepSeek-Coder-6.7B-Instruct on Code-290k-ShareGPT, using Alpaca instruction format. Provider: ajibawa-2023. Access: Open.
Aggregate Rating 1200 ± 37, rank #9617 of 10916 rated models, from 7 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EvalPlus (HumanEval+ & MBPP+) | 59.7 | Pass@1 avg (%) | 57.3 |
| Open LLM Leaderboard v1 - GSM8K | 3.49 | Accuracy (%) (5-shot) | 21.6 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 41.95 | MC2 (%) (0-shot) | 18.8 |
| Open LLM Leaderboard v1 - MMLU | 34.89 | Accuracy (%) (5-shot) | 18.2 |
| Open LLM Leaderboard v1 - ARC Challenge | 34.9 | Normalized accuracy (%) (25-shot) | 12.8 |
| Open LLM Leaderboard v1 - HellaSwag | 51.99 | Normalized accuracy (%) (10-shot) | 11 |
| Open LLM Leaderboard v1 - WinoGrande | 52.64 | Accuracy (%) (5-shot) | 6.9 |
Interactive version: theaggregate.ai/model?slug=code-290k-6-7b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-08.