CodeLlama-7B: benchmark results
Provider: Meta.
Aggregate Rating 1250 ± 40, rank #9256 of 10916 rated models, from 16 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| BizBench - FinCode | 10.7 | Accuracy (%; 121 CFA and CPA exam questions answered by gene | 53.3 |
| BizBench - CodeFinQA | 34 | Accuracy (%; 844 FinQA questions over report text and tables | 46.7 |
| BizBench - CodeTAT-QA | 70.9 | Accuracy (%; 392 TAT-QA table questions answered by generati | 46.7 |
| BizBench | 52.1 | Mean accuracy (%) over the eight BizBench tasks (FinKnow, Fo | 35.7 |
| EvalPlus (HumanEval+ & MBPP+) | 41.1 | Pass@1 avg (%) | 34.7 |
| BizBench - FormulaEval | 52 | Accuracy (%; 50 financial-formula functions to complete from | 33.3 |
| BizBench - ConvFinQA Extract | 82.4 | Accuracy (%; 916 ConvFinQA questions answerable by extractin | 20 |
| BizBench - SEC-Num | 53.1 | Accuracy (%; 2,000 SEC 10-K and 10-Q filing snippets, extrac | 20 |
| BizBench - TAT-QA Extract | 77.4 | Accuracy (%; 248 TAT-QA questions answerable by extracting a | 20 |
| BigCode Models Leaderboard | 30 | HumanEval Python Pass@1 (%) | 18.6 |
| Big Code Memorization - HumanEval pass@50 | 22.11 | HumanEval pass@50 (%) | 5.9 |
| Big Code Memorization - HumanEval-ET pass@50 | 18.32 | HumanEval-ET pass@50 (%) | 5.9 |
Interactive version: theaggregate.ai/model?slug=codellama-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.