Horangi 4 - GLP - Semantic Analysis: leaderboard
Semantic analysis category of the Horangi Korean LLM Leaderboard 4 (W&B Korea): the mean of KoBALT-700 semantics and HAE-RAE Bench reading comprehension accuracy. 0-100, higher is better.
Metric: Score (%). Source: horangi.ai. Saturation forecast: Around February 2027. 105 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) (High) | 90.5 | #64 (Gemini 3.1 Pro (Preview)) |
| 2 | Claude Fable 5 (High) | 90 | #20 (Claude Fable 5) |
| 3 | GPT-5.6 Sol (Max) | 88.5 | #22 (GPT-5.6 Sol) |
| 4 | Gemini 3 Pro (Preview) (High) | 88 | #71 (Gemini 3 Pro (Preview)) |
| 5 | Gemini 3.5 Flash | 87.91 | #66 |
| 6 | GPT-5.5 (xHigh) | 87.5 | #33 (GPT-5.5) |
| 7 | GPT-5.6 Terra (Max) | 87 | #40 (GPT-5.6 Terra) |
| 8 | GPT-5.6 Luna (Max) | 86.5 | #79 (GPT-5.6 Luna) |
| 9 | Qwen 3.7 Max | 86.08 | #85 |
| 10 | GPT-5.2 (2025-12-11) (xHigh) | 86 | |
| 11 | GLM-5 | 85.5 | #149 |
| 12 | Grok 4.5 | 85.31 | #54 |
| 13 | Claude Opus 4.6 (High) | 85 | #67 (Claude Opus 4.6) |
| 14 | Claude Opus 4.5 (20251101) (High) | 85 | #81 (Claude Opus 4.5 (20251101)) |
| 15 | GPT-5 | 84.5 | #105 |
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-semantic-analysis · How It Works · Data refreshed daily, snapshot 2026-11-05.