Back to Models
Google Deep MindConditional Open

TranslateGemma 27B

TranslateGemma 27B is a translation-specialized foundation model developed by Google DeepMind. It has a parameter scale of approximately 29B and supports a long context window of up to 128K.

Parameters

29.0B

Context Window

128K

License

Gemma License

Release Date

2026-01-15

API Pricing

API pricing for this model is not yet available

Strengths

  • High performance specialized in translation
  • 128K broad context window
  • Developed by Google DeepMind

Weaknesses

  • Huge 55GB file size
  • Unknown adaptability to general tasks
  • Subject to specific license restrictions

Use Cases

  • High-quality multilingual translation
  • Translating long documents
  • Foundation for translation-specific AI

Deep Analysis

Parameters

27B

Largest and most capable TranslateGemma variant

Performance

Highest quality in family

Best translation quality among all TranslateGemma models

Languages

55 core / ~500 extended

Most comprehensive language coverage

Release Date

January 15, 2026

Part of TranslateGemma family launch

Multimodal

Yes

Best image text translation in the family

Deployment

Server / GPU required

Requires significant compute for deployment

Training

SFT + RL

Distilled from Gemini with reinforcement learning optimization

Strengths

  • Highest translation quality in the TranslateGemma family
  • Best multimodal (image text) translation performance
  • 55 core languages with broadest extended coverage
  • Distilled from Gemini models for maximum fidelity
  • Open weights for full customization
  • Suitable for the most demanding translation tasks

Weaknesses

  • Largest model, requires significant GPU resources
  • Not suitable for mobile or edge deployment
  • Marginal quality gain over 12B may not justify 2x compute cost
  • Translation-focused, not general-purpose
  • Extended language pairs lack confirmed evaluation metrics

Competitor Comparison

ModelArenaSWEGPQAPrice
TranslateGemma 27BN/AN/AN/AFree (open weights)
TranslateGemma 12BN/AN/AN/AFree (open weights)
Gemma 3 27B (baseline)~1430N/A~70%Free (open weights)
NLLB-200 (54.5B)N/AN/AN/AFree (open weights)
GPT-5 (general)~1480~80%~90%$5/$20 per 1M

TranslateGemma 27B is the largest and most capable variant in Google's open translation model family. It delivers the highest translation quality across 55 languages with multimodal image text translation support. Best suited for demanding server-side translation workloads.

Analysis generated: 2026-05-24