Horangi 4 - GLP - General Knowledge: leaderboard
General knowledge category of the Horangi Korean LLM Leaderboard 4 (W&B Korea): the mean of KMMLU and HAE-RAE Bench (without reading comprehension) accuracy. 0-100, higher is better.
Metric: Score (%). Source: horangi.ai. Saturation forecast: Estimated already saturated. 105 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) (High) | 97 | #64 (Gemini 3.1 Pro (Preview)) |
| 2 | GPT-5.6 Sol (Max) | 96.49 | #22 (GPT-5.6 Sol) |
| 3 | Gemini 3.5 Flash | 96.43 | #66 |
| 4 | GPT-5.5 (xHigh) | 96 | #33 (GPT-5.5) |
| 5 | GPT-5.6 Terra (Max) | 95.49 | #40 (GPT-5.6 Terra) |
| 6 | Gemini 3 Pro (Preview) (High) | 95.49 | #71 (Gemini 3 Pro (Preview)) |
| 7 | GPT-5.6 Luna (Max) | 94.49 | #79 (GPT-5.6 Luna) |
| 8 | GPT-5.2 (2025-12-11) (xHigh) | 94.49 | |
| 9 | Claude Fable 5 (High) | 93.99 | #20 (Claude Fable 5) |
| 10 | Claude Opus 4.8 (High) | 93.48 | #39 (Claude Opus 4.8) |
| 11 | Claude Sonnet 4.6 (High) | 92.97 | #94 (Claude Sonnet 4.6) |
| 12 | Claude Opus 4.5 (20251101) (High) | 92.97 | #81 (Claude Opus 4.5 (20251101)) |
| 13 | Motif 3 | 92.71 | #195 |
| 14 | GPT-5.4 (2026-03-05) (xHigh) | 92.47 | #63 (GPT-5.4 (2026-03-05)) |
| 15 | Grok 4.5 | 92.45 | #54 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=horangi-4-glp-general-knowledge · How It Works · Data refreshed daily, snapshot 2026-11-05.