Benchmarks

Language performance

How well does each AI model perform when prompted in different languages? Scores are based on multilingual benchmark prompts written in native 6 languages.

Tests run weekly · Models with image/audio/TTS capabilities excluded

🇬🇧

English

1 benchmark runs · English

#	Model	Score	Runs
1	Claude Fable 5Anthropic	100.0	1
2	gpt-oss-120bOVH AI Endpoints (GRA)	100.0	3
3	Mistral-Small-3.2-24B-Instruct-2506OVH AI Endpoints (GRA)	100.0	2
4	Claude Opus 4.8Anthropic	99.5	2
5	Gemini 3.1 Flash LiteGoogle Gemini	99.5	2

🇳🇱

Nederlands

1 benchmark runs · Dutch

#	Model	Score	Runs
1	Claude Fable 5Anthropic	100.0	1
2	Claude Opus 4.5Anthropic	98.3	3
3	Claude Opus 4.6Anthropic	98.0	3
4	Gemma 4 26B A4B ITGoogle Gemini	96.0	3
5	Gemma 4 31B ITGoogle Gemini	95.0	3

🇩🇪

Deutsch

2 benchmark runs · German

#	Model	Score	Runs
1	Claude Haiku 4.5Anthropic	100.0	2
2	Nano Banana ProGoogle Gemini	100.0	2
3	gpt-3.5-turboOpenAI	100.0	2
4	gpt-4o-search-previewOpenAI	100.0	2
5	Qwen2.5-VL-72B-InstructOVH AI Endpoints (GRA)	100.0	1

🇫🇷

Français

2 benchmark runs · French

#	Model	Score	Runs
1	Claude Opus 4.6Anthropic	100.0	2
2	Claude Opus 4Anthropic	100.0	2
3	Gemini Flash-Lite LatestGoogle Gemini	100.0	2
4	Gemini 3.1 Flash LiteGoogle Gemini	99.5	2
5	Claude Sonnet 4.5Anthropic	99.5	2

🇪🇸

Español

2 benchmark runs · Spanish

#	Model	Score	Runs
1	gpt-4o-2024-08-06OpenAI	100.0	2
2	gpt-4.1OpenAI	100.0	3
3	gpt-4o-miniOpenAI	100.0	2
4	gpt-4o-mini-2024-07-18OpenAI	100.0	3
5	gpt-4.1-mini-2025-04-14OpenAI	100.0	2

🇹🇷

Türkçe

1 benchmark runs · Turkish

#	Model	Score	Runs
1	gpt-oss-120bOVH AI Endpoints (GRA)	100.0	1
2	gpt-oss-20bOVH AI Endpoints (GRA)	100.0	1
3	gpt-4o-miniOpenAI	100.0	1
4	gpt-4.1-mini-2025-04-14OpenAI	100.0	1
5	Meta-Llama-3_3-70B-InstructOVH AI Endpoints (GRA)	100.0	1