What are the hardware requirements for Swahili Gemma?

Swahili Gemma requires approximately 17.1 GB of VRAM and supports a context window of 8K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Swahili Gemma and how large is it?

Swahili Gemma is developed by Mollel, a model with 7b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Swahili Gemma perform on standard benchmarks?

Swahili Gemma has the following published scores: MMLU 61.02. Compare against reference models on this page or on the LLM Explorer leaderboards.

Swahili Gemma by Mollel — VRAM 17.1GB, 8K context

Name: Swahili Gemma
Author: Mollel

Swahili Gemma is an open-source language model by Mollel. Features: 7b LLM, VRAM: 17.1GB, Context: 8K, License: apache-2.0, Quantized, HF Score: 61.3, LLM Explorer Score: 0.15, Arc: 59, HellaSwag: 76.4, MMLU: 61, TruthfulQA: 52.1, WinoGrande: 75.6, GSM8K: 43.8.

4bit Conversational En Endpoints compatible Gemma Pytorch Quantized Region:us Sharded Trl Unsloth

Model Card on HF 🤗: https://huggingface.co/Mollel/Swahili_Gemma

Swahili Gemma Benchmarks

ARC: 58.96 vs 96.7 (so35)^-39%

HellaSwag: 76.4 vs 95.3 (gpt4)^-19.8%

MMLU: 61.02 vs 88.3 (so35)^-30.9%

TruthfulQA: 52.1 vs 59 (gpt4)^-11.7%

WinoGrande: 75.61 vs 87.5 (gpt4)^-13.6%

GSM8K: 43.82 vs 96.4 (so35)^-54.5%

LLME Score: 0.14536

^nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

What is the LLM Explorer Rank (Score)

🌟 Advertise your project 🚀

Swahili Gemma Parameters and Internals

Model Type

text-generation-inference, transformers, unsloth, gemma, trl

LLM Name	Swahili Gemma
Repository 🤗	https://huggingface.co/Mollel/Swahili_Gemma
Base Model(s)	gemma-7b-bnb-4bit /gemma-7b-bnb-4bit
Model Size	7b
Required VRAM	17.1 GB
Updated	2026-04-23
Maintainer	Mollel
Model Type	gemma
Model Files	5.0 GB: 1-of-4 5.0 GB: 2-of-4 5.0 GB: 3-of-4 2.1 GB: 4-of-4
Supported Languages	en
Quantization Type	4bit
Model Architecture	GemmaForCausalLM
License	apache-2.0
Context Length	8192
Model Max Length	8192
Transformers Version	4.38.2
Tokenizer Class	GemmaTokenizer
Padding Token	<pad>
Vocabulary Size	256000
Torch Data Type	float16

Quantized Models of the Swahili Gemma

Model	Likes	Downloads	VRAM
DAD Model Gemma V3	0	15	5 GB

Best Alternatives to Swahili Gemma

Best Alternatives	Context / RAM	Downloads	Likes
...t Cleaner Gemma 32k Merged 16b	31K / 17.1 GB	5	0
...ruct Malayalam Model Vllm 4bit	16K / 5.6 GB	7	0
Gemma 1.1 7B It 4bit	8K / 5.9 GB	186	4
Gemma 1.1 7B It 4bit	8K / 5.9 GB	3	1
Gemma 7B 8bit	8K / 9.4 GB	15	1
... Codegemma 2 7B It Alpaca V1.2	8K / 17.1 GB	16	0
Codegemma 7B It 8bit	8K / 9.5 GB	170	5
Fbt Gemma 7B	8K / 17.1 GB	5	0
Codegemma 7B 4bit	8K / 5.9 GB	43	1
Codegemma 7B It 4bit	8K / 5.9 GB	28	1

Note: green Score (e.g. "73.2") means that the model is better than Mollel/Swahili_Gemma.

Rank the Swahili Gemma Capabilities

🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

What open-source LLMs or SLMs are you in search of? 53254 in total.

Email us: info@extractum.io. Our Privacy Policy | Terms and Conditions | Suggest an improvement.

Our Social Media →

Original data from HuggingFace, OpenCompass and various public git repos.

Check out Ag3ntum — our secure, self-hosted AI agent for server management.

Release v20260328a

Support LLM Explorer

Swahili Gemma by Mollel