KorT - Korean-English Translation: leaderboard
KorT (kort.worldsw.dev): about 100 private bidirectional Korean-English items across 28 phenomenon types. Every pair of system outputs is compared in both orders by GPT-5.6 Sol and Gemini 3.1 Pro judges with a reference, fitted to Bradley-Terry Elo. Higher is better; Elo is relative within this record. Commercial and research MT engines are excluded.
Metric: Elo (pairwise LLM-judged). Source: kort.worldsw.dev. Saturation forecast: Around July 2027. 15 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Luna | 1201.8 | #79 |
| 2 | Gemini 3.6 Flash | 1179.3 | #57 |
| 3 | GPT-5.6 Terra | 1178.8 | #40 |
| 4 | Gemini 3.7 Flash | 1141.3 | #24 |
| 5 | Gemini 3.1 Pro (Preview) | 1132.3 | #64 |
| 6 | GLM-5.2 | 1119.1 | #80 |
| 7 | GPT-5.4 Mini | 1114.5 | #228 |
| 8 | Nemotron 3 Ultra | 1103.7 | #245 |
| 9 | DeepSeek V4 Pro (0813) | 1101.7 | #58 |
| 10 | DeepSeek V4 Flash (0731) | 1096.1 | #84 |
| 11 | Gemini 3.5 Flash Lite | 1072 | #174 |
| 12 | MiniMax-M3 | 1040.5 | #151 |
| 13 | Gemma 4 31B | 1034.4 | #212 |
| 14 | Nemotron 3 Super | 950.2 | #344 |
| 15 | GPT-OSS-120B | 926.3 | #346 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=kort-korean-english-translation · How It Works · Data refreshed daily, snapshot 2026-11-05.