internlm2-chat-20B: benchmark results
Shanghai AI Lab's 20B InternLM2 chat model with 200K context, tool use and code-interpreter support, from the January 2024 InternLM2 release. Provider: Shanghai AI Lab. Released 2024-01-17. Access: Open.
Unified ELO 1477 ± 1, rank #800 of 1392 rated models, from 33 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open CoT - LSAT Logical Reasoning | 22.35 | CoT Gain (%) | 97.7 |
| Open CoT - LSAT Reading Comprehension | 24.16 | CoT Gain (%) | 97.7 |
| Open CoT - LogiQA 2 | 17.62 | CoT Gain (%) | 97.7 |
| Open CoT Leaderboard | 16.11 | Average CoT Gain (%) | 97.7 |
| SALAD-Bench Base | 98.15 | Safety Score (%) | 97 |
| Open CoT - LogiQA | 8.15 | CoT Gain (%) | 90.5 |
| Open CoT - LSAT Analytical Reasoning | 8.26 | CoT Gain (%) | 89.3 |
| Open Chinese LLM - ARC Challenge | 59.39 | Accuracy (%) | 84.1 |
| Open Chinese LLM - HellaSwag | 66.29 | Accuracy (%) | 80.7 |
| Enkrypt AI - Jailbreak Risk | 4.34 | Risk Score | 78.8 |
| Enkrypt AI - Safety Risk | 21.72 | Risk Score | 75.8 |
| Open Chinese LLM - WinoGrande | 65.35 | Accuracy (%) | 72.1 |
Interactive version: theaggregate.ai/model?slug=internlm2-chat-20b · How It Works · Data refreshed daily, snapshot 2026-09-05.