KorSET
Models respond to Korean prompts adapted to local language, law, and culture. A model judge checks whether each response violates the relevant safety policy.
Selectstar, which ran three annual safety challenges, created KorSET. Its public card lists 9,249 prompts. The authors gate downloads and report five target models.
- Made by
- DATUMO AI Safety Team
- Kind
- Deception and alignment
- Level
- Everyday
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Propensity
- Risk
- Sexual content, Hate, Specialized advice, Privacy, Non-violent crimes, Violent crimes, Cyber offence
- Behavior
- Harmful use, Bias
- Language
- Korean
- Interaction
- Single turn
- Items from
- Generated, Another benchmark
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Content safety
- Strongest model tested
-
Llama 3.1-8B #203 on the capability index
- Last model added
- March 2025
- Scores refreshed
- November 20th, 2026
gemma2-9B-it tops the leaderboard with 19.4% EXAONE-3.5-7.8B-Instruct leads with 19.9% among models released by March 2025
Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. The headline is the plain mean of the five risk-category rates in Table 3. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is EXAONE-3.5-7.8B-Instruct, at 19.9%. This view shows only models we track, each at its best.
The newest model with a score here came out in March 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1 gemma2-9B-it Direct 19.4%
- 2
EXAONE-3.5-7.8B-Instruct GCG 19.9%
- 3
Gemma 3 12B GCG 20%
- 4
Qwen2.5-7B GCG 22.3%
- 5
Llama 3.1-8B GPTFuzzer 42.3%
Source: The KorSET paper
- 1
EXAONE-3.5-7.8B-Instruct GCG 19.9%
- 2
Gemma 3 12B GCG 20%
- 3
Qwen2.5-7B GCG 22.3%
- 4
Llama 3.1-8B GPTFuzzer 42.3%
Source: The KorSET paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | gemma2-9B-it | 19.4% | Direct | Official leaderboard | |
| 2 | | 19.9% | GCG | Dec 2024 | Official leaderboard |
| 3 | | 20% | GCG | Mar 2025 | Official leaderboard |
| 4 | | 22.3% | GCG | Sep 2024 | Official leaderboard |
| 5 | | 42.3% | GPTFuzzer | Jul 2024 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 19.9% | GCG | Dec 2024 | Official leaderboard |
| 2 | | 20% | GCG | Mar 2025 | Official leaderboard |
| 3 | | 22.3% | GCG | Sep 2024 | Official leaderboard |
| 4 | | 42.3% | GPTFuzzer | Jul 2024 | Official leaderboard |
Rankings by part
We read 5 parts from the benchmark’s paper, each ranking models on one part of the test, such as Toxic Language, Unfair Representation and Misinformation Harms. EXAONE-3.5-7.8B-Instruct leads 2 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Toxic Language
4 models we track have a score here.
- 1
Qwen2.5-7B ↑2 7.65%
- 2
EXAONE-3.5-7.8B-Instruct ↓1 7.73%
- 3
Gemma 3 12B ↓1 13.5%
- 4
Llama 3.1-8B 29.5%
Unfair Representation
4 models we track have a score here.
- 1
EXAONE-3.5-7.8B-Instruct 24.5%
- 2
Llama 3.1-8B ↑2 28.4%
- 3
Gemma 3 12B ↓1 28.5%
- 4
Qwen2.5-7B ↓1 32.5%
Misinformation Harms
4 models we track have a score here.
- 1
Gemma 3 12B ↑1 12.3%
- 2
EXAONE-3.5-7.8B-Instruct ↓1 13.8%
- 3
Qwen2.5-7B 18.6%
- 4
Llama 3.1-8B 44.7%
Information & Safety Harms
4 models we track have a score here.
- 1
EXAONE-3.5-7.8B-Instruct 6.65%
- 2
Gemma 3 12B 9.75%
- 3
Qwen2.5-7B 15.7%
- 4
Llama 3.1-8B 53.6%
Malicious Use
4 models we track have a score here.
- 1
Gemma 3 12B ↑1 15.1%
- 2
Qwen2.5-7B ↑1 21.1%
- 3
EXAONE-3.5-7.8B-Instruct ↓2 33.4%
- 4
Llama 3.1-8B 41.5%
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: The KorSET paper
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| GPTFuzzer | 45.2% | ||||
| GCG | 44.8% | ||||
| GPTFuzzer | 44% | ||||
| Direct | 43.6% | ||||
| AutoDAN | 43.3% | ||||
| GPTFuzzer | 43.1% | ||||
| TAP | 42.6% | ||||
| GPTFuzzer | 42.3% | ||||
| AutoDAN | 36.3% | ||||
| AutoDAN | 35.1% | ||||
| AutoDAN | 33.6% | ||||
| TAP | 32.2% | ||||
| TAP | 27% | ||||
| Direct | 24.4% | ||||
| TAP | 24.1% | ||||
| Direct | 22.6% | ||||
| GCG | 22.3% | ||||
| Direct | 21.4% | ||||
| GCG | 20% | ||||
| GCG | 19.9% |
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: The KorSET paper
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| GPTFuzzer | 45.2% | ||||
| GCG | 44.8% | ||||
| GPTFuzzer | 44% | ||||
| Direct | 43.6% | ||||
| AutoDAN | 43.3% | ||||
| GPTFuzzer | 43.1% | ||||
| TAP | 42.6% | ||||
| GPTFuzzer | 42.3% | ||||
| AutoDAN | 36.3% | ||||
| AutoDAN | 35.1% | ||||
| AutoDAN | 33.6% | ||||
| TAP | 32.2% | ||||
| TAP | 27% | ||||
| Direct | 24.4% | ||||
| TAP | 24.1% | ||||
| Direct | 22.6% | ||||
| GCG | 22.3% | ||||
| Direct | 21.4% | ||||
| GCG | 20% | ||||
| GCG | 19.9% |
3 models have held number one in the benchmark’s paper since 2023
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The KorSET paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official KorSET leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Safety and alignment
- FORTRESS In use Paired adversarial and benign national-security misuse prompts. Rubric-graded risk score on adversarial and over-refusal score on benign prompts. Claude Opus 5.5 5.78%
- StrongREJECT In use Forbidden-prompt jailbreak benchmark with a rubric grader that scores how useful a non-refusal actually is.
Inkling 98.6%
- Gray Swan IPI Arena benchmark In use Indirect prompt-injection attacks from public red-teaming competitions replayed on tool, coding and browser agents. Attack success rate at k attempts.
Gemini 4 Carbon 0%
- Covert misbehavior (Thinking Machines) In use Agentic tasks that check whether a model misbehaves and conceals it. Scored by the rate of covert misbehavior.
Inkling 2 Mini 8.30%
- Destructive actions (Thinking Machines) In use Agentic tasks that check whether a model takes destructive actions it was not authorized to take. Scored by the rate of destructive actions.
Inkling 2 Small 4.70%
- Misbehavior (Thinking Machines) In use Agentic tasks that check whether a model fabricates results, ignores instructions or hides what it did when a task is hard. Scored by the rate of misbehavior.
Inkling 2 Mini 9.80%
Data audit
8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 20 months ago, in March 2025, for Gemma 3 12B. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
- Pass Independent runs Scores from the benchmark’s paper (5 entries).
- Pass Models scored 4 models we track, plus 1 other entry on its leaderboard.
- Warning Newest score The newest came 20 months ago, in March 2025, for Gemma 3 12B.
- Note Frontier models None of the top 10 has a score. The strongest tested is Llama 3.1-8B (#203).
- Note Headroom Lower is better, so there’s no ceiling to reach.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 20 rows for models we track and 1 other entry at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.