OpenAssistant-SFT-7-Llama-30B-HF: benchmark results
TheBloke's Hugging Face publication of OpenAssistant's LLaMA 30B SFT 7, whose original release is OpenAssistant/oasst-sft-7-llama-30b-xor. Conversion publication is credited separately from the original model creator. The card specifies reconstruction from OpenAssistant XOR weights and original Llama weights. Provider: TheBloke. Access: Open.
Aggregate Rating 1360 ± 15, rank #7454 of 10916 rated models, from 68 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MMLU-by-task - Global Facts | 42 | Accuracy (%) | 91.7 |
| MMLU-by-task - Professional Medicine | 60.66 | Accuracy (%) | 91.1 |
| MMLU-by-task - International Law | 79.34 | Accuracy (%) | 90.2 |
| MMLU-by-task - High School Statistics | 49.54 | Accuracy (%) | 90.1 |
| MMLU-by-task - High School Biology | 71.61 | Accuracy (%) | 89.6 |
| MMLU-by-task - Marketing | 85.9 | Accuracy (%) | 89.5 |
| MMLU-by-task - Conceptual Physics | 51.49 | Accuracy (%) | 89 |
| MMLU-by-task - College Physics | 35.29 | Accuracy (%) | 88.7 |
| MMLU-by-task - High School Microeconomics | 63.03 | Accuracy (%) | 88.7 |
| MMLU-by-task - Econometrics | 36.84 | Accuracy (%) | 88.3 |
| MMLU-by-task - High School Psychology | 79.08 | Accuracy (%) | 88.2 |
| MMLU-by-task - Professional Psychology | 61.6 | Accuracy (%) | 88 |
Interactive version: theaggregate.ai/model?slug=openassistant-sft-7-llama-30b-hf · How It Works · Data refreshed daily, snapshot 2026-11-08.