CodeLlama-13B-Instruct: benchmark results
Provider: Meta. Access: Open.
Aggregate Rating 1312 ± 38, rank #8503 of 10916 rated models, from 27 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MERA Code - YABLoCo | 8.17 | pass@1 (%) | 66.9 |
| EvoEval Tool Use | 54 | Pass@1 (%) | 59 |
| BigCode Models Leaderboard | 50.6 | HumanEval Python Pass@1 (%) | 57.6 |
| EvoEval Creative | 25 | Pass@1 (%) | 51 |
| EvoEval Difficult | 21 | Pass@1 (%) | 47 |
| EvoEval | 37.16 | Pass@1 (%) | 46 |
| EvoEval Subtle | 46 | Pass@1 (%) | 44 |
| EvoEval Concise | 48.17 | Pass@1 (%) | 42 |
| EvoEval Verbose | 46.34 | Pass@1 (%) | 41 |
| EvoEval Combine | 8 | Pass@1 (%) | 39 |
| CyberSecEval | 37.27 | Average Injection Success Rate (imported) | 33.3 |
| MERA Code - JavaTestGen | 9.25 | pass@1 (%) | 33.1 |
Interactive version: theaggregate.ai/model?slug=codellama-13b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-08.