SEA-HELM (Indonesian) - Summarization: leaderboard
Abstractive summarization of Indonesian news articles. Normalized 0-100 score from AI Singapore's SEA-HELM leaderboard: (score - random baseline) / (maximum - baseline), floored at 0 and averaged over eight runs; higher is better.
Metric: Normalized Score. Source: leaderboard.sea-lion.ai. Saturation forecast: Estimated already saturated. 58 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | MERaLiON-2-10B | 19.28 | #600 |
| 2 | Llama-SEA-LION-v3-70B-IT | 19.14 | #407 |
| 3 | Llama 3.3 70B Instruct | 18.79 | #541 |
| 4 | Llama 4 Scout Instruct | 18.15 | #547 |
| 5 | Llama 3.1 8B Instruct | 17.95 | #1017 |
| 6 | Gemma 4 26B A4B (IT) (Reasoning) | 17.74 | #293 (Gemma 4 26B A4B (IT)) |
| 7 | Qwen 3.5 27B (Reasoning) | 17.25 | #192 (Qwen 3.5 27B) |
| 8 | Llama-SEA-LION-v3-8B-IT | 17.15 | #800 |
| 9 | Apertus-8B-Instruct-2509 | 16.97 | #1028 |
| 10 | Llama 3.2 3B Instruct | 16.9 | #1327 |
| 11 | Qwen 3.5 122B A10B (Reasoning) | 16.85 | #184 (Qwen 3.5 122B A10B) |
| 12 | Qwen 3.6 35B A3B (Reasoning) | 16.62 | #238 (Qwen 3.6 35B A3B) |
| 13 | Qwen 3 14B (Reasoning) | 16.56 | #542 (Qwen 3 14B) |
| 14 | Qwen 3.5 9B (Reasoning) | 16.5 | #391 (Qwen 3.5 9B) |
| 15 | GLM-4.7 Flash (Reasoning) | 16.41 | #543 (GLM-4.7 Flash) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=sea-helm-indonesian-summarization · How It Works · Data refreshed daily, snapshot 2026-11-05.