Icelandic LLM - WikiQA: leaderboard
Metric: LLM Judge Score (%). Source: huggingface.co. 24 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Fable 5.1 | 73.95 |
| 2 | Claude Opus 5 | 69.08 |
| 3 | Gemini 3.1 Pro (Preview) | 68.92 |
| 4 | GPT-6 | 66.71 |
| 5 | Gemini 3.7 Flash | 66.58 |
| 6 | Gemini 3.8 Flash | 66.55 |
| 7 | Gemini 3.5 Flash | 66.37 |
| 8 | GPT-5.6 Sol | 64.5 |
| 9 | Claude Opus 4.8 | 62.74 |
| 10 | Gemini 2.5 Pro | 56.18 |
| 11 | Claude Sonnet 5 | 49.87 |
| 12 | GPT-5.6 Luna | 47.84 |
| 13 | GPT-5.6 Terra | 47.32 |
| 14 | DeepSeek V4 Pro | 46.05 |
| 15 | Kimi K3 | 46.05 |
Interactive version: theaggregate.ai/benchmark?slug=icelandic-llm-wikiqa · How It Works · Data refreshed daily, snapshot 2026-09-19.