Skip to content

Model qwen3 30b

hydropix edited this page Dec 23, 2025 · 1 revision

qwen3:30b

Ollama Model ID: qwen3:30b


Summary

Metric Value
Average Score 🟠 6.7/10
Accuracy 7.4/10
Fluency 6.7/10
Style 6.4/10
Languages Tested 19
Total Translations 95
Best Language Chinese (Traditional) (8.0)
Worst Language Bengali (5.2)

Language Performance

Top Languages

Rank Language Overall Accuracy Fluency Style
1 Chinese (Traditional) 🟡 8.0 8.0 7.8 7.6
2 Chinese (Simplified) 🟡 7.8 7.8 8.0 7.2
3 Italian 🟡 7.6 8.0 7.6 7.2
4 German 🟡 7.4 7.8 7.0 7.4
5 Portuguese 🟡 7.4 8.4 7.4 7.4
6 French 🟡 7.2 7.8 7.2 6.8
7 Spanish 🟡 7.2 7.6 7.4 6.6
8 Russian 🟡 7.2 7.6 7.4 6.6
9 Arabic 🟡 7.0 7.6 7.0 6.4
10 Ukrainian 🟠 6.8 7.8 6.8 6.8
View all 19 languages
Rank Language Overall Accuracy Fluency Style
1 Chinese (Traditional) 🟡 8.0 8.0 7.8 7.6
2 Chinese (Simplified) 🟡 7.8 7.8 8.0 7.2
3 Italian 🟡 7.6 8.0 7.6 7.2
4 German 🟡 7.4 7.8 7.0 7.4
5 Portuguese 🟡 7.4 8.4 7.4 7.4
6 French 🟡 7.2 7.8 7.2 6.8
7 Spanish 🟡 7.2 7.6 7.4 6.6
8 Russian 🟡 7.2 7.6 7.4 6.6
9 Arabic 🟡 7.0 7.6 7.0 6.4
10 Ukrainian 🟠 6.8 7.8 6.8 6.8
11 Polish 🟠 6.6 7.4 6.4 6.4
12 Vietnamese 🟠 6.6 7.4 6.6 6.6
13 Thai 🟠 6.4 7.4 6.4 6.2
14 Hindi 🟠 6.4 7.2 6.4 6.2
15 Japanese 🟠 6.0 6.8 6.2 5.6
16 Korean 🟠 6.0 7.0 6.0 5.6
17 Tamil 🟠 5.6 6.6 5.6 5.2
18 Hebrew 🟠 5.6 6.2 5.6 5.0
19 Bengali 🟠 5.2 6.2 5.2 4.8

Performance by Category

Asian Languages

Language Overall Accuracy Fluency Style
Chinese (Traditional) 🟡 8.0 8.0 7.8 7.6
Chinese (Simplified) 🟡 7.8 7.8 8.0 7.2
Vietnamese 🟠 6.6 7.4 6.6 6.6
Thai 🟠 6.4 7.4 6.4 6.2
Hindi 🟠 6.4 7.2 6.4 6.2
Japanese 🟠 6.0 6.8 6.2 5.6
Korean 🟠 6.0 7.0 6.0 5.6
Tamil 🟠 5.6 6.6 5.6 5.2
Bengali 🟠 5.2 6.2 5.2 4.8

Category Average: 🟠 6.4

European Major Languages

Language Overall Accuracy Fluency Style
Italian 🟡 7.6 8.0 7.6 7.2
German 🟡 7.4 7.8 7.0 7.4
Portuguese 🟡 7.4 8.4 7.4 7.4
French 🟡 7.2 7.8 7.2 6.8
Spanish 🟡 7.2 7.6 7.4 6.6
Polish 🟠 6.6 7.4 6.4 6.4

Category Average: 🟡 7.2

Cyrillic Languages

Language Overall Accuracy Fluency Style
Russian 🟡 7.2 7.6 7.4 6.6
Ukrainian 🟠 6.8 7.8 6.8 6.8

Category Average: 🟡 7.0

Semitic Languages

Language Overall Accuracy Fluency Style
Arabic 🟡 7.0 7.6 7.0 6.4
Hebrew 🟠 5.6 6.2 5.6 5.0

Category Average: 🟠 6.3


Translation Examples

Best Translation

Chinese (Traditional) - Walden Score: 🟢 9.0/10

View Translation

Original:

Translation:

我步入森林,是為了以最清醒的意識生活,直面生命的本質,探求它所要教導的真理,以免臨終之際才驚覺自己未曾真正活過。我不願過非真實的生活,因生命如此珍貴;亦不願屈服,除非萬不得已。我渴望深入生活,吸盡生命的精髓。

Feedback: The translation captures Thoreau's philosophical intent and existential urgency with remarkable fidelity. The rendering of key concepts like "live deliberately" as "以最清醒的意識生活" and "suck out all the marrow of life" as "吸盡生命的精髓" demonstrates sophisticated understanding of both languages and the source material's deeper meanings. Minor stylistic variations exist (e.g., "屈服" for "practise resignation" is slightly more literal than interpretive), but the overall effect preserves Thoreau's passionate,

Worst Translation

Bengali - Moby-Dick Score: 🔴 4.0/10

View Translation

Original:

Translation:

আমাকে ইশ্মায়েল বলুন। কিছু বছর আগে—কোন মাপে ঠিক করা নেই—আমার পকেটে কিছু টাকা নেই, বরং কোনও বিশেষ কাজ নেই যা আমাকে শুকনো ভূমির উপর আকর্ষণ করে, তাই আমি ভেবেছিলাম, একটু জলপথে ঘুরে বেড়াব এবং পৃথিবীর জলাভূমি অংশটুকু দেখব। এটাই আমার অভ্যাস, ক্ষুধা দূর করে আর রক্ত প্রবাহ নিয়ন্ত্রণ করার জন্য। যখনই আমার মুখ কঠিন হয়ে যায়; যখনই আমার আত্মায় ভিজে-ভিজে নভেম্বরের মতো আবহাওয়া ছড়িয়ে পড়ে, তখন আমি বুঝি, সময় এসে গেছে যত দ্রুত সম্ভব সমুদ্রে যাওয়ার।

Feedback: The translation contains significant meaning distortions and awkward phrasings that undermine comprehension. Key errors include 'ক্ষুধা' (hunger) instead of 'melancholy/spleen,' 'জলাভূমি' (wetland) instead of 'watery,' and grammatically incorrect constructions that make the text difficult to parse, failing to capture Melville's contemplative, archaic tone.


Score Distribution

Score Range Count Percentage
🟢 Excellent (9-10) 1 1.1%
🟡 Good (7-8) 55 57.9%
🟠 Acceptable (5-6) 37 38.9%
🔴 Poor (3-4) 2 2.1%
⚫ Failed (1-2) 0 0.0%

Performance Metrics

Metric Value
Average Translation Time 46571.0ms
Success Rate 100.0%

← Back to Home | All Models

Clone this wiki locally