Skip to content

Model translategemma 12b

hydropix edited this page Jan 16, 2026 · 1 revision

translategemma:12b

Ollama Model ID: translategemma:12b


Summary

Metric Value
Average Score 🟠 6.8/10
Accuracy 7.2/10
Fluency 6.8/10
Style 6.1/10
Languages Tested 19
Total Translations 95
Best Language Spanish (7.8)
Worst Language Bengali (5.6)

Language Performance

Top Languages

Rank Language Overall Accuracy Fluency Style
1 Spanish 🟡 7.8 7.8 7.8 7.0
2 Ukrainian 🟡 7.4 7.6 7.2 6.8
3 Portuguese 🟡 7.2 7.6 7.4 6.4
4 Chinese (Simplified) 🟡 7.2 7.2 7.2 6.2
5 Chinese (Traditional) 🟡 7.2 7.2 7.2 6.2
6 Arabic 🟡 7.2 7.4 7.2 6.4
7 French 🟡 7.0 7.0 7.2 6.0
8 Italian 🟡 7.0 7.4 7.2 6.2
9 Russian 🟡 7.0 7.8 7.0 6.8
10 German 🟠 6.8 7.4 6.6 6.4
View all 19 languages
Rank Language Overall Accuracy Fluency Style
1 Spanish 🟡 7.8 7.8 7.8 7.0
2 Ukrainian 🟡 7.4 7.6 7.2 6.8
3 Portuguese 🟡 7.2 7.6 7.4 6.4
4 Chinese (Simplified) 🟡 7.2 7.2 7.2 6.2
5 Chinese (Traditional) 🟡 7.2 7.2 7.2 6.2
6 Arabic 🟡 7.2 7.4 7.2 6.4
7 French 🟡 7.0 7.0 7.2 6.0
8 Italian 🟡 7.0 7.4 7.2 6.2
9 Russian 🟡 7.0 7.8 7.0 6.8
10 German 🟠 6.8 7.4 6.6 6.4
11 Vietnamese 🟠 6.6 7.4 6.6 6.2
12 Thai 🟠 6.6 7.4 6.6 6.4
13 Hindi 🟠 6.6 7.2 6.6 6.0
14 Polish 🟠 6.4 6.8 6.6 5.6
15 Hebrew 🟠 6.4 7.0 6.4 6.0
16 Japanese 🟠 6.2 7.0 6.2 5.8
17 Tamil 🟠 6.2 7.0 6.2 5.6
18 Korean 🟠 6.0 6.8 6.0 5.4
19 Bengali 🟠 5.6 6.6 5.6 5.4

Performance by Category

European Major Languages

Language Overall Accuracy Fluency Style
Spanish 🟡 7.8 7.8 7.8 7.0
Portuguese 🟡 7.2 7.6 7.4 6.4
French 🟡 7.0 7.0 7.2 6.0
Italian 🟡 7.0 7.4 7.2 6.2
German 🟠 6.8 7.4 6.6 6.4
Polish 🟠 6.4 6.8 6.6 5.6

Category Average: 🟡 7.0

Cyrillic Languages

Language Overall Accuracy Fluency Style
Ukrainian 🟡 7.4 7.6 7.2 6.8
Russian 🟡 7.0 7.8 7.0 6.8

Category Average: 🟡 7.2

Asian Languages

Language Overall Accuracy Fluency Style
Chinese (Simplified) 🟡 7.2 7.2 7.2 6.2
Chinese (Traditional) 🟡 7.2 7.2 7.2 6.2
Vietnamese 🟠 6.6 7.4 6.6 6.2
Thai 🟠 6.6 7.4 6.6 6.4
Hindi 🟠 6.6 7.2 6.6 6.0
Japanese 🟠 6.2 7.0 6.2 5.8
Tamil 🟠 6.2 7.0 6.2 5.6
Korean 🟠 6.0 6.8 6.0 5.4
Bengali 🟠 5.6 6.6 5.6 5.4

Category Average: 🟠 6.5

Semitic Languages

Language Overall Accuracy Fluency Style
Arabic 🟡 7.2 7.4 7.2 6.4
Hebrew 🟠 6.4 7.0 6.4 6.0

Category Average: 🟠 6.8


Translation Examples

Best Translation

Spanish - Pride and Prejudice Score: 🟢 9.0/10

View Translation

Original:

Translation:

Es una verdad universalmente reconocida que un hombre soltero, en posesión de una buena fortuna, debe estar buscando una esposa. Por muy desconocidos que sean los sentimientos o las opiniones de un hombre así al establecerse en un vecindario, esta verdad está tan arraigada en la mente de las familias vecinas que se considera que él es la propiedad legítima de una u otra de sus hijas.

Feedback: The translation faithfully preserves Austen's meaning and ironic tone, with excellent word choices like 'universalmente reconocida' and 'propiedad legítima' that capture the original's satirical voice. Minor stylistic variations exist, but the translation reads naturally in Spanish while maintaining the formal, period-appropriate register of the source text.

Worst Translation

French - A Study in Scarlet Score: 🟠 5.0/10

View Translation

Original:

Translation:

"Vous avez été en Afghanistan, je crois", ai-je observé. "Comment avez-vous pu le savoir ?", ai-je demandé, stupéfait. "Ce n'est pas grave", a-t-il répondu en riant, "la question est maintenant celle de l'hémoglobine. Ne voyez-vous pas l'importance de cette découverte que j'ai faite ?" "C'est certainement intéressant d'un point de vue chimique", ai-je répondu, "mais en termes pratiques..." "Voyez-vous, mon ami, c'est la découverte médico-légale la plus pratique depuis des années. Ne comprenez-vo

Feedback: The translation contains a critical error: the first line incorrectly attributes Holmes's observation to Watson ("ai-je observé"), reversing the speaker. Additionally, the dialogue attribution becomes confused, and the translation loses the distinctive voice and deductive brilliance of Holmes's character, making it read more formally and less naturally than the original's conversational tone.


Score Distribution

Score Range Count Percentage
🟢 Excellent (9-10) 1 1.1%
🟡 Good (7-8) 59 62.1%
🟠 Acceptable (5-6) 35 36.8%
🔴 Poor (3-4) 0 0.0%
⚫ Failed (1-2) 0 0.0%

Performance Metrics

Metric Value
Average Translation Time 1950.0ms
Success Rate 100.0%

← Back to Home | All Models

Clone this wiki locally