Skip to content

Benchmark

UzLiB overall accuracy (all items)

Multilingual benchmarks compare model performance across languages. Higher scores are better.

Models measured
74
Scores recorded
74
Sources
1
IQ weight
1.26

Score distribution

0.562
Mean
0.548
Median
0.154
Std dev
0.375
10th percentile
0.795
90th percentile
0.233
Min
0.2330.852
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.26 for chat models, measured over 73 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51Qwen3 235B A22B Instruct 2507temperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.487
52Claude Sonnet 4.5temperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.485
53Gemma 3 12B ITtemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.468
54GPT OSS 20btemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.458
55Qwen3 32Btemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.456
56Qwen3.5 27Btemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.454
57Gemma 3n E4B ITtemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.437
58Qwen3.5 9Btemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.431
59Llama 3.3 70B Instructtemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.429
60Qwen3.5 122B-A10Btemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.42
61C4AI Command A 03 2025temperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.419
62Qwen3.5 397B A17Btemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.411
63Mistral Nemotemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.41
64Qwen 3.8 27Btemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.396
65Qwen3 8Btemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.385
66Qwen3.5 35B A3Btemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.38
67Llama 3.1 8B Instructtemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.375
68Qwen3 14Btemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.374
69Qwen3 4Btemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.345
70Phi 4temperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.34
71Mistral (7B) Instruct v0.3temperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.325
72Llama 4 Mavericktemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.281
73Llama 3.2 3Btemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.267
74Llama 3.2 1B Instructtemperature: 1.0, top_p: 0.95· UzLiB Uzbek linguistic benchmark data docs0.233

Showing 51–74 of 74 results.

Data credits: UzLiB, Tahrirchi (Abror Shopulatov). Public LEADERBOARD.md numbers; repository MIT covers software and documentation, no separate results licence stated and none asserted.