Title: Gemma 4 31B IT vs Gemma 4 31B: Parameetrid, Litsents %E2%80%94 LLM EU
Canonical: https://llmeu.com/et/compare/gemma-4-31b-it-vs-gemma-4-31b/
Markdown mirror: https://llmeu.com/et/compare/gemma-4-31b-it-vs-gemma-4-31b.md
Machine-readable: https://llmeu.com/llms.txt (index) - https://llmeu.com/llms-full.txt (full catalogue) - https://llmeu.com/llms-catalogue.json

# Gemma 4 31B IT vs Gemma 4 31B

Aseta kaks kuni neli mudelit k%C3%B5rvuti v%C3%A4ljadel, mis m%C3%A4%C3%A4ravad kasutuselev%C3%B5tu: kontekst, hind, piirkond, litsents, keeled, majutamine ja see, mis on kontrollitud.

Kui kaks mudelit erinevad, n%C3%A4itab see tabel erinevust ja l%C3%B5petab. LLM EU ei avalda v%C3%B5itjam%C3%A4rki ilma selle %C3%BClesande dokumenteeritud skoorita ning puuduv n%C3%A4itaja j%C3%A4%C3%A4b puuduma, mitte ei muutu nulliks.

## Millal kasutada

### [Gemma 4 31B IT](/et/models/gemma-4-31b-it/)

#### Tugevused

- Apache-2.0 licence, unlike earlier Gemma generations
- 256K-token context window
- text and image input with variable resolution
- vendor reports multilingual coverage in over 140 languages

#### Piirangud

- all published scores are self-reported by the vendor
- 31B parameters need a large-memory accelerator at BF16
- no evaluation has been run by LLM EU

#### %C3%9Clesanded

[N%C3%A4gemine](/et/tasks/vision/), [Arutlusk%C3%A4ik](/et/tasks/reasoning/), [Mitmekeelne tugi](/et/tasks/multilingual-support/), [Kodeerimine](/et/tasks/coding/), [Kokkuv%C3%B5tete tegemine](/et/tasks/summarization/)

### [Gemma 4 31B](/et/models/gemma-4-31b/)

#### Tugevused

- open weights are published on Hugging Face and can be downloaded without a vendor agreement
- Apache-2.0 licence, so commercial self-hosting is permitted without a separate agreement

#### Piirangud

- 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 32.7B parameters, self-hosting needs a high-memory 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

#### %C3%9Clesanded

[Mitmekeelne tugi](/et/tasks/multilingual-support/)

**%C3%9Clesanded:** [Mitmekeelne tugi](/et/tasks/multilingual-support/)

## Mudelite v%C3%B5rdlemine

- [Command R (08-2024) vs Gemma 4 31B IT](/et/compare/command-r-08-2024-vs-gemma-4-31b-it/)
- [DeepSeek Coder V2 Lite Instruct vs Gemma 4 31B IT](/et/compare/deepseek-coder-v2-lite-instruct-vs-gemma-4-31b-it/)
- [DeepSeek-R1-Distill-Qwen-32B vs Gemma 4 31B IT](/et/compare/deepseek-r1-distill-qwen-32b-vs-gemma-4-31b-it/)
- [DeepSeek V2 Lite vs Gemma 4 31B IT](/et/compare/deepseek-v2-lite-vs-gemma-4-31b-it/)
- [Devstral Small 2 (24B) vs Gemma 4 31B IT](/et/compare/devstral-small-2-24b-vs-gemma-4-31b-it/)
- [Diffusiongemma 26B A4B It vs Gemma 4 31B IT](/et/compare/diffusiongemma-26b-a4b-it-vs-gemma-4-31b-it/)