Skip to content

Benchmark

Horangi 4 subset: ko_hallulens_nonexistent

Higher scores are better.

Models measured
79
Scores recorded
83
Sources
1
IQ weight
0.13

Score distribution

0.691
Mean
0.75
Median
0.203
Std dev
0.33
10th percentile
0.93
90th percentile
0.22
Min
0.220.99
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.13 for chat models, measured over 79 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Gemini 3.1 Pro Previewgemini-3.1-pro-preview: high-effort· Horangi Korean Leaderboard 4 data docs0.99
2Qwen3.7 Maxqwen3.7-max· Horangi Korean Leaderboard 4 data docs0.98
3GLM 5.2glm-5.2: high-effort· Horangi Korean Leaderboard 4 data docs0.96
4Qwen3 4B Instruct 2507Qwen3-4B-Instruct-2507· Horangi Korean Leaderboard 4 data docs0.96
5Gemini 3.5 Flashgemini-3.5-flash· Horangi Korean Leaderboard 4 data docs0.95
6GLM 5 Turboglm-5-turbo· Horangi Korean Leaderboard 4 data docs0.95
7GPT-5 minigpt-5-mini-2025-08-07· Horangi Korean Leaderboard 4 data docs0.95
8Qwen 3 235B A22B Thinking 2507qwen3-235b-a22b-thinking-2507· Horangi Korean Leaderboard 4 data docs0.949
9Qwen3.5 397B A17Bqwen3.5-397b-a17b· Horangi Korean Leaderboard 4 data docs0.93
10Qwen3.7 Plusqwen3.7-plus· Horangi Korean Leaderboard 4 data docs0.929
11Gemma 4 31B ITgemma-4-31B-it: enable-thinking· Horangi Korean Leaderboard 4 data docs0.91
12Gemma 4 31B ITGemma-4-31B-IT-NVFP4: enable-thinking· Horangi Korean Leaderboard 4 data docs0.9
13GPT 4ogpt-4o-2024-11-20· Horangi Korean Leaderboard 4 data docs0.88
14Qwen3 0.6BQwen3-0.6B· Horangi Korean Leaderboard 4 data docs0.88
15Qwen3.5 122B-A10Bqwen3.5-122b-a10b· Horangi Korean Leaderboard 4 data docs0.88
16Claude Sonnet 5claude-sonnet-5: high-effort· Horangi Korean Leaderboard 4 data docs0.86
17Claude Opus 4.8claude-opus-4-8: high-effort· Horangi Korean Leaderboard 4 data docs0.85
18GLM 5glm-5· Horangi Korean Leaderboard 4 data docs0.85
19Grok 4 Fastgrok-4-fast-reasoning· Horangi Korean Leaderboard 4 data docs0.85
20Gemma 4 E4B-itgemma-4-E4B-it: enable-thinking· Horangi Korean Leaderboard 4 data docs0.84
21Qwen3 1.7BQwen3-1.7B· Horangi Korean Leaderboard 4 data docs0.83
22Qwen3.5 35B A3Bqwen3.5-35b-a3b· Horangi Korean Leaderboard 4 data docs0.83
23Qwen3.5 9Bqwen3.5-9b· Horangi Korean Leaderboard 4 data docs0.827
24GPT 5.1gpt-5.1-2025-11-13· Horangi Korean Leaderboard 4 data docs0.82
25GPT-4.1 nanogpt-4.1-nano· Horangi Korean Leaderboard 4 data docs0.82
26Kimi K2.5kimi-k2.5· Horangi Korean Leaderboard 4 data docs0.82
27Qwen3.5 27Bqwen3.5-27b· Horangi Korean Leaderboard 4 data docs0.82
28DeepSeek Reasonerdeepseek-v3.2: high-effort· Horangi Korean Leaderboard 4 data docs0.81
29Qwen3 14BQwen3-14B· Horangi Korean Leaderboard 4 data docs0.81
30Qwen3 32BQwen3-32B· Horangi Korean Leaderboard 4 data docs0.81
31Gemma 4 E2B-itgemma-4-E2B-it: enable-thinking· Horangi Korean Leaderboard 4 data docs0.79
32GLM 4.7glm-4.7· Horangi Korean Leaderboard 4 data docs0.78
33Qwen3.6 Flashqwen3.6-flash· Horangi Korean Leaderboard 4 data docs0.78
34Claude Opus 4claude-opus-4-20250514: high-effort· Horangi Korean Leaderboard 4 data docs0.77
35Nemotron 3 Super 120B A12BNVIDIA-Nemotron-3-Super-120B-A12B-FP8· Horangi Korean Leaderboard 4 data docs0.77
36Qwen3.6 Plus Previewqwen3.6-plus-preview· Horangi Korean Leaderboard 4 data docs0.77
37Grok 4.1 Fastgrok-4-1-fast-reasoning· Horangi Korean Leaderboard 4 data docs0.76
38Nemotron 3 Nano 30B A3BNVIDIA-Nemotron-3-Nano-30B-A3B-BF16· Horangi Korean Leaderboard 4 data docs0.76
39Qwen3 4BQwen3-4B· Horangi Korean Leaderboard 4 data docs0.76
40Gemini 3 Pro Previewgemini-3-pro-preview: high-effort· Horangi Korean Leaderboard 4 data docs0.75
41Grok 4.5grok-4.5· Horangi Korean Leaderboard 4 data docs0.75
42Mistral Smallmistral-small-2603· Horangi Korean Leaderboard 4 data docs0.75
43MiMo V2 Promimo-v2-pro· Horangi Korean Leaderboard 4 data docs0.74
44Claude Haiku 4.5claude-haiku-4-5-20251001: high-effort· Horangi Korean Leaderboard 4 data docs0.73
45Grok 4.1 Fastgrok-4-1-fast-non-reasoning· Horangi Korean Leaderboard 4 data docs0.73
46Claude Opus 4.1claude-opus-4-1-20250805· Horangi Korean Leaderboard 4 data docs0.72
47GLM 5V Turboglm-5v-turbo· Horangi Korean Leaderboard 4 data docs0.72
48Claude Opus 4.5claude-opus-4-5-20251101: high-effort· Horangi Korean Leaderboard 4 data docs0.711
49Claude Sonnet 4.5claude-sonnet-4-5-20250929: high-effort· Horangi Korean Leaderboard 4 data docs0.71
50Gemma 4 26B A4Bgemma-4-26B-A4B-it: enable-thinking· Horangi Korean Leaderboard 4 data docs0.71

Showing 1–50 of 83 results.

Data credits: Horangi 4, Weights & Biases Korea; public publisher-selected numeric evaluation results