Title: Gemma 3 4B It vs Gemma 3 27B It: Parametre, Licens %E2%80%94 LLM EU
Canonical: https://llmeu.com/da/compare/gemma-3-4b-it-vs-gemma-3-27b-it/
Markdown mirror: https://llmeu.com/da/compare/gemma-3-4b-it-vs-gemma-3-27b-it.md
Machine-readable: https://llmeu.com/llms.txt (index) - https://llmeu.com/llms-full.txt (full catalogue) - https://llmeu.com/llms-catalogue.json

# Gemma 3 4B It vs Gemma 3 27B It

S%C3%A6t to til fire modeller side om side p%C3%A5 de felter, der afg%C3%B8r en udrulning: kontekst, pris, region, licens, sprog, hosting og hvad der er verificeret.

Hvor to modeller adskiller sig, viser denne tabel forskellen og stopper. LLM EU offentligg%C3%B8r ikke et vinderbadge uden en dokumenteret score for den opgave, og et manglende tal forbliver manglende i stedet for at blive et nul.

## Hvorn%C3%A5r man bruger den

### [Gemma 3 4B It](/da/models/gemma-3-4b-it/)

#### Styrker

- open weights are published on Hugging Face and can be downloaded without a vendor agreement
- at 4.3B parameters it fits on a single accelerator
- instruction-tuned checkpoint with a chat template in the repository

#### Begr%C3%A6nsninger

- catalogue only %E2%80%94 LLM EU does not run this model, and nothing on this card implies an LLM EU endpoint
- no evaluation has been run by LLM EU
- the published configuration records no context window, so the card carries none
- at 4.3B parameters, the weights fit on a single modest accelerator
- vision checkpoints are heavier to serve than their parameter count alone suggests, because the image encoder and pre-processing run alongside the language model

#### Opgaver

[Flersproget support](/da/tasks/multilingual-support/), [Vision](/da/tasks/vision/), [Lav pris](/da/tasks/cheap/)

### [Gemma 3 27B It](/da/models/gemma-3-27b-it/)

#### Styrker

- open weights are published on Hugging Face and can be downloaded without a vendor agreement
- instruction-tuned checkpoint with a chat template in the repository

#### Begr%C3%A6nsninger

- catalogue only %E2%80%94 LLM EU does not run this model, and nothing on this card implies an LLM EU endpoint
- no evaluation has been run by LLM EU
- the published configuration records no context window, so the card carries none
- at 27.4B parameters, self-hosting needs a single accelerator with enough memory for the weights
- vision checkpoints are heavier to serve than their parameter count alone suggests, because the image encoder and pre-processing run alongside the language model

#### Opgaver

[Flersproget support](/da/tasks/multilingual-support/), [Vision](/da/tasks/vision/), [R%C3%A6sonnement](/da/tasks/reasoning/)

**Opgaver:** [Flersproget support](/da/tasks/multilingual-support/), [Vision](/da/tasks/vision/)

## Sammenlign modeller

- [Aya 23 8B vs Gemma 3 4B It](/da/compare/aya-23-8b-vs-gemma-3-4b-it/)
- [Aya Expanse 8B vs Gemma 3 4B It](/da/compare/aya-expanse-8b-vs-gemma-3-4b-it/)
- [Blip2 Flan T5 Xl vs Gemma 3 4B It](/da/compare/blip2-flan-t5-xl-vs-gemma-3-4b-it/)
- [Blip2 Opt 2.7B vs Gemma 3 4B It](/da/compare/blip2-opt-2-7b-vs-gemma-3-4b-it/)
- [Bloom 3B vs Gemma 3 4B It](/da/compare/bloom-3b-vs-gemma-3-4b-it/)
- [Command R (08-2024) vs Gemma 3 27B It](/da/compare/command-r-08-2024-vs-gemma-3-27b-it/)