Best AI for translation

As of 2026-07-18, Anthropic: Claude Opus 4.8 (Fast) is the strongest model for this task in our daily benchmark, where over 350 AI models are evaluated and the best are presented.

Translation is where the gap between models is widest for Scandinavian languages. A model can write excellent English and still produce stiff, English-flavoured Norwegian with the wrong word order and obvious anglicisms.

For translation we rank by how well the model actually lands in the target language. Ranked by language score, from today’s benchmark.

Top choice

Anthropic: Claude Opus 4.8 (Fast)

anthropic

Score: 8.3/10

Language: 9.3/10

Speed: 131 t/s

$10.00

Runner-up

Anthropic: Claude Haiku 4.5

anthropic

Score: 8.0/10

Language: 9.3/10

Speed: 113 t/s

$1.00

Third choice

Google: Gemma 3 4B

google

Score: 8.0/10

Language: 9.3/10

Speed: 237 t/s

$0.05

Today’s ranking

Model	Provider	Score	Language	Speed	Input price/1M
Anthropic: Claude Opus 4.8 (Fast)	anthropic	8.3/10	9.3/10	131 t/s	$10.00
Anthropic: Claude Haiku 4.5	anthropic	8.0/10	9.3/10	113 t/s	$1.00
Google: Gemma 3 4B	google	8.0/10	9.3/10	237 t/s	$0.05
Cohere: Command R+ (08-2024)	cohere	7.5/10	9.3/10	160 t/s	$2.50
Perplexity: Sonar Pro	perplexity	7.7/10	9.0/10	68 t/s	$3.00
Meta: Llama 3.1 8B Instruct	meta-llama	5.2/10	9.0/10	392 t/s	$0.05
Mistral: Mistral Small 3	mistralai	7.8/10	8.7/10	234 t/s	$0.05
IBM: Granite 4.0 Micro	ibm-granite	5.4/10	8.3/10	179 t/s	$0.02

What separates a good translator from a bad one?

Weak models translate word by word. The result is grammatically correct but instantly recognisable as machine output: English sentence structure, wrong prepositions and phrases no native speaker would use. Strong models rewrite the sentence the way a native writer would, and hit the right register — formal in a letter, conversational in a chat.

When should you not use AI for translation?

Legal agreements, medical information and marketing copy intended for publication should be reviewed by a human. AI is excellent for first drafts, internal communication and understanding text in a language you do not read — but it can be confidently wrong, and the error is hard to spot precisely because the text reads well.

hvilkenAI is independent. We have no affiliate agreements, sponsored placements or commercial ties to the AI providers we test. Every recommendation rests on test results alone.

Frequently asked questions

Which AI is best at translating into Norwegian?

It changes from month to month. The ranking on this page comes from today’s benchmark, where we evaluate over 350 models and present the ones that perform strongest on the language.

Is AI better than Google Translate?

For running text the best language models are clearly better, because they account for context, tone and register. Google Translate remains fast and practical for single words and short sentences.

Can I translate sensitive documents with AI?

Not without checking the provider’s data processing terms. Personal data and trade secrets should only go to services with an approved data processing agreement.

Why are Scandinavian languages harder than English for AI?

Models are trained on far more English text. Norwegian, Swedish and Danish are small languages in the training data, which is exactly why the differences between models show up so clearly here.

Other tasks

Doing something else? See which model is best for it.

Writing and answering email faster.

Job applications and CVs

Cover letters and CVs that avoid generic phrasing.

Summaries

Boiling long text down to what matters.

Meeting notes

Minutes and action items from raw notes.

Proofreading

Fixing language without changing the content.

Customer service

Chat and email replies fast enough for real time.

→ Other tasks