SeaLLM-7B-v2.5: benchmark results
Alibaba DAMO's Gemma-7B-based SeaLLM (April 2024) fine-tuned for Southeast Asian languages such as Vietnamese, Thai, and Indonesian. Provider: Alibaba. Released 2024-04-10. Access: Open.
Aggregate Rating 1436 ± 13, rank #1071 of 1656 rated models, from 191 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - GSM8K | 74.37 | Accuracy (%) (5-shot) | 99 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 56.71 | Exact match, flexible extract (%) | 90.6 |
| Open Korean LLM Leaderboard | 42.38 | Average Score (%) | 86.4 |
| Open Ko-LLM Leaderboard | 42.38 | Average (%) | 86.2 |
| Open Arabic LLM - Alghafa Multiple Choice Rating Sentiment NO Neutral Task | 80.69 | Accuracy (%) | 85.8 |
| Open Arabic LLM - Arabic MMLU Arabic Language (Grammar) | 50.41 | Accuracy (%) | 80.6 |
| Open LLM Leaderboard v1 - MMLU | 64.93 | Accuracy (%) (5-shot) | 78.8 |
| Open Arabic LLM - Arabic MMLU HT Abstract Algebra | 36 | Accuracy (%) | 76.9 |
| Open Arabic LLM - Alghafa Multiple Choice Grounded Statement Soqal Task | 91.33 | Accuracy (%) | 76.5 |
| Open Arabic LLM - Madinah QA Arabic Language (Grammar) | 50.14 | Accuracy (%) | 75.9 |
| Open Arabic LLM - Arabic MMLU HT Electrical Engineering | 57.24 | Accuracy (%) | 74.1 |
| Open Arabic LLM - Madinah QA Average | 57.99 | Accuracy (%) | 71 |
Interactive version: theaggregate.ai/model?slug=seallm-7b-v2-5 · How It Works · Data refreshed daily, snapshot 2026-10-20.