Gemma 1.1 7B (IT): benchmark results
Google's Gemma 1.1 7B instruction-tuned model, released under the Gemma license; the linked model card provides no further architecture or training details. Provider: Google. Released 2024-04-05. Access: Open.
Aggregate Rating 1357 ± 13, rank #1368 of 1656 rated models, from 126 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - KorNAT-Harmless | 67.58 | Normalized accuracy (%) | 90 |
| ChineseSafe Benchmark | 71.7 | Accuracy (%) | 83.3 |
| Open Korean LLM - Ko-IFEval | 39.27 | Strict accuracy, prompt/instruction mean (%) | 80.7 |
| SLM Trustworthiness - Machine Ethics | 82.19 | TrustLLM machine-ethics accuracy (%; sub-task mean) | 80 |
| SLM Trustworthiness - Privacy | 60.13 | TrustLLM privacy accuracy (%; sub-task mean) | 80 |
| SLM Trustworthiness - Robustness | 68.9 | TrustLLM robustness accuracy (%; sub-task mean) | 80 |
| SLM Trustworthiness - TrustLLM Overall | 63.99 | TrustLLM score (%; mean of ethics, privacy, robustness and f | 80 |
| Open Korean LLM - Ko-GPQA-Diamond | 24.75 | Normalized accuracy (%) | 69.5 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 32.9 | Exact match, flexible extract (%) | 66.7 |
| Open Korean LLM - KorNAT-CKA | 26.43 | Normalized accuracy (%) | 64.4 |
| Open LLM Leaderboard - MuSR | 42.3 | Score | 63.3 |
| CyberMetric | 75.4 | 500-question accuracy (%) | 62.5 |
Interactive version: theaggregate.ai/model?slug=gemma-1-1-7b-it · How It Works · Data refreshed daily, snapshot 2026-10-20.