# Korean Benchmarks Leaderboard (2026)

> Cross-market leaderboard comparing global and Korean-market AI models on Korean benchmark rows including KMMLU, KMMLU-Hard, CLIcK, and KoBALT.

Canonical page: https://benchlm.ai/leaderboards/korean-benchmarks

## Current Leader

[Solar Open](/models/solar-open-2) leads this view with an average Korean benchmark score of 87.1.

## Rankings

| Rank | Model | Creator | Type | KMMLU | KMMLU-Hard | KMMLU-Pro | CLIcK | KoBALT | Avg |
|------|-------|---------|------|-------|------------|-----------|-------|--------|-----|
| 1 | [Solar Open](/models/solar-open-2) | Upstage | Global | N/A | N/A | 78.4 | 90.7 | N/A | 87.1 |
| 2 | [A.X](/models/a-x-k2) | SK Telecom | Global | N/A | N/A | 80.5 | 91.6 | 73 | 81.7 |
| 3 | [Solar](/models/solar-pro-4) | Upstage | Global | N/A | N/A | 79.2 | N/A | N/A | 79.2 |
| 4 | [K-EXAONE 2.0](/models/k-exaone-2-0) | LG AI Research | Global | N/A | N/A | 69.1 | 84.2 | N/A | 76.7 |
| 5 | [Kanana 2](/models/kanana-2-3b-instruct) | Kakao | Global | 43.3 | N/A | N/A | N/A | N/A | 43.3 |
| 6 | [Kanana 2](/models/kanana-2-1-3b-instruct) | Kakao | Global | 42.8 | N/A | N/A | N/A | N/A | 42.8 |

## How to Use This Page

Use this mixed view when you need to compare Korean-market specialists against global frontier models on the same Korean benchmark rows.

- [Regional-only Korean LLM leaderboard](/leaderboards/korean-llm)
- [KMMLU guide](/guides/kmmlu-explained)
- [Best Korean LLM guide](/best/korean-llm)