internlm2-base-7B: benchmark results
Shanghai AI Lab's InternLM2 base 7B model, the foundation checkpoint in the InternLM2 family intended for further domain adaptation, supporting contexts up to 200,000 characters. Provider: Shanghai AI Lab. Released 2024-01-17. Access: Open.
Aggregate Rating 1353 ± 28, rank #1376 of 1656 rated models, from 40 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Chinese LLM - HellaSwag | 63.75 | Accuracy (%) | 74.3 |
| Open Chinese LLM - WinoGrande | 65.43 | Accuracy (%) | 72.8 |
| Open Chinese LLM - ARC Challenge | 52.56 | Accuracy (%) | 49.9 |
| Open PL LLM - PSC (generative, 5-shot) | 85.26 | Binary F1 (%) | 41.9 |
| Open PL LLM - KLEJ NER (multiple choice, 5-shot) | 42.52 | Accuracy (%) | 40 |
| Open PL LLM - CBD (multiple choice, 5-shot) | 18.73 | Macro F1 (%) | 35 |
| Open Portuguese LLM - BLUEX | 39.22 | Accuracy (%) | 32.2 |
| Open Portuguese LLM - ENEM | 44.51 | Accuracy (%) | 31.6 |
| Open Portuguese LLM - FaQuAD NLI | 49.57 | Macro F1 (%) | 31.5 |
| Open PL LLM - DYK (multiple choice, 5-shot) | 34.38 | Binary F1 (%) | 30.2 |
| Open Portuguese LLM - ASSIN2 RTE | 78.83 | Macro F1 (%) | 29.8 |
| Open PL LLM - Multiple Choice | 20.46 | Average Multiple-Choice Score (%) | 29.7 |
Interactive version: theaggregate.ai/model?slug=internlm2-base-7b · How It Works · Data refreshed daily, snapshot 2026-10-20.