MPT-30B-Chat: benchmark results
MosaicML's MPT-30B-Chat, a chat fine-tune of the MPT-30B decoder-style transformer pretrained on 1 trillion tokens of English text and code, with an 8K ALiBi-extended context. Provider: Databricks. Released 2023-06-22.
Aggregate Rating 1314 ± 12, rank #1431 of 1635 rated models, from 94 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MMLU-by-task - College Mathematics | 38 | Accuracy (%) | 89.5 |
| MMLU-by-task - TruthfulQA MC2 | 52.42 | Accuracy (%) | 86.7 |
| MMLU-by-task - HellaSwag | 63.53 | Accuracy (%) | 83.5 |
| MMLU-by-task - Abstract Algebra | 34 | Accuracy (%) | 82 |
| MMLU-by-task - High School Mathematics | 30.37 | Accuracy (%) | 79.4 |
| MMLU-by-task - High School US History | 76.47 | Accuracy (%) | 79.3 |
| MMLU-by-task - Business Ethics | 56 | Accuracy (%) | 77.9 |
| MMLU-by-task - College Medicine | 52.02 | Accuracy (%) | 77.9 |
| MMLU-by-task - College Physics | 31.37 | Accuracy (%) | 77.6 |
| MMLU-by-task - Global Facts | 37 | Accuracy (%) | 76.3 |
| MMLU-by-task - TruthfulQA MC1 | 34.15 | Accuracy (%) | 76.3 |
| MMLU-by-task - College Biology | 59.03 | Accuracy (%) | 74.7 |
Interactive version: theaggregate.ai/model?slug=mpt-30b-chat · How It Works · Data refreshed daily, snapshot 2026-10-15.