Hebrew LLM - Summarization (Pairwise): leaderboard
The Summarization task of the Hebrew LLM Leaderboard — Hebrew text summarization judged in pairwise comparisons; pairwise score (%).
Metric: Pairwise Score (%). Source: huggingface.co. Saturation forecast: Around January 2027. 34 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.1 | 65.98 | #135 |
| 2 | GLM-5.3 Flash | 60.52 | #65 |
| 3 | MiniMax-M2.7 | 56.51 | #258 |
| 4 | GPT-5 Mini | 55.11 | #186 |
| 5 | Claude Sonnet 4.5 | 55.07 | #142 |
| 6 | Gemini 3 Pro (Preview) | 51.76 | #66 |
| 7 | Gemini 3 Flash (Preview) | 49.75 | #82 |
| 8 | Gemini 2.5 Pro | 49.1 | #152 |
| 9 | MiniMax-M2.5 | 48.8 | #303 |
| 10 | Gemma 4 31B (IT) | 48.5 | #176 |
| 11 | GLM-5.1 | 47.9 | #97 |
| 12 | Qwen 3.5 397B A17B | 47.34 | #146 |
| 13 | Gemini 2.5 Flash | 46.94 | #245 |
| 14 | Qwen 3.5 122B A10B | 46.19 | #173 |
| 15 | GLM-4.6 | 46.19 | #252 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=hebrew-llm-summarization-pairwise · How It Works · Data refreshed daily, snapshot 2026-10-20.