Horangi 4 - Ko-MT-Bench: leaderboard
Metric: Judge rating (1-10, x10). Source: horangi.ai. 105 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.6 (High) | 90.13 |
| 2 | GPT-5.1 | 89.63 |
| 3 | Gemini 3.5 Flash | 89.63 |
| 4 | Claude Opus 4.5 (20251101) (High) | 89.63 |
| 5 | Gemini 3 Flash (Preview) | 89.62 |
| 6 | Claude Sonnet 4.6 (High) | 89.5 |
| 7 | Claude Sonnet 4.5 (High) | 89.38 |
| 8 | GPT-OSS-120B (High) | 89.25 |
| 9 | Gemini 3.1 Pro (Preview) (High) | 89.25 |
| 10 | Claude Sonnet 5 (High) | 89.25 |
| 11 | Qwen 3.7 Max | 89.13 |
| 12 | Gemini 3 Pro (Preview) (High) | 89.13 |
| 13 | Gemini 3.1 Flash Lite (Preview) | 89 |
| 14 | MiMo-V2-Pro | 89 |
| 15 | GPT-5.2 (xHigh) | 88.88 |
Interactive version: theaggregate.ai/benchmark?slug=horangi-4-ko-mt-bench · How It Works · Data refreshed daily, snapshot 2026-09-19.