CodeLlama-70B-Instruct: benchmark results
Provider: Meta. Access: Open.
Aggregate Rating 1397 ± 37, rank #6271 of 10916 rated models, from 36 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CyberSecEval | 12.93 | Average Injection Success Rate (imported) | 100 |
| EvoEval Tool Use | 65 | Pass@1 (%) | 92 |
| BigCode Models Leaderboard | 75.6 | HumanEval Python Pass@1 (%) | 89.8 |
| MERA Code - YABLoCo | 12.02 | pass@1 (%) | 89.2 |
| EvoEval Creative | 41 | Pass@1 (%) | 81 |
| EvoEval Subtle | 65 | Pass@1 (%) | 80 |
| EvoEval | 51.66 | Pass@1 (%) | 70 |
| EvoEval Combine | 18 | Pass@1 (%) | 70 |
| EvoEval Difficult | 31 | Pass@1 (%) | 70 |
| EvoEval Verbose | 65.85 | Pass@1 (%) | 69 |
| EvalPlus (HumanEval+ & MBPP+) | 65.9 | Pass@1 avg (%) | 67.7 |
| BigCodeBench | 40.7 | Pass@1 (%) | 62.8 |
Interactive version: theaggregate.ai/model?slug=codellama-70b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-08.