vicuna-7B-v1.3: benchmark results
LMSYS's 7B chat assistant, fine-tuned from LLaMA on about 125,000 ShareGPT conversations under a non-commercial research license. Provider: LMSYS. Released 2023-06-18. Access: Open.
Aggregate Rating 1282 ± 10, rank #1507 of 1656 rated models, from 111 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| HELM Classic - CivilComments | 62.04 | Exact Match (%) | 86.4 |
| HELM Classic - LegalSupport | 60.53 | Exact Match (%) | 85.3 |
| MMLU-by-task - College Mathematics | 36 | Accuracy (%) | 80.2 |
| LLM Trustworthy - Adversarial | 52.16 | Trust Score (%) | 79.2 |
| HELM Classic - RAFT | 69.32 | Exact Match (%) | 77.3 |
| MMLU-by-task - High School Statistics | 45.83 | Accuracy (%) | 76.9 |
| HELM Classic - QuAC | 39.22 | F1 (%) | 72.3 |
| HELM Classic - Entity Matching | 84.84 | Exact Match (%) | 71.2 |
| HELM Classic - TruthfulQA | 29.2 | Exact Match (%) | 71.2 |
| MMLU-by-task - Elementary Mathematics | 32.54 | Accuracy (%) | 70.2 |
| HELM Classic - MMLU | 43.36 | Exact Match (%) | 69.7 |
| HELM Classic - GSM8K | 13.4 | Exact Match (%) | 69.1 |
Interactive version: theaggregate.ai/model?slug=vicuna-7b-v1-3 · How It Works · Data refreshed daily, snapshot 2026-10-20.