LLM EXPLORER
 LLM Explorer: A Curated Large Language Model Directory and Analytics
Dark Theme
  LLM  List    LLM Hosting    LLM Leaderboards    Blog    Newsfeed    Advertise

Aya Expanse 8B by CohereForAI

 »  All LLMs  »  CohereForAI  »  Aya Expanse 8B   URL Share it on

  32413   
  Arxiv:2406.18682   Arxiv:2407.02552   Arxiv:2408.14960   Arxiv:2410.10801   Arxiv:2412.04261   Ar   Autotrain compatible   Cohere   Conversational   Cs   De   El   En   Es   Fa   Fr   He   Hi   Id   It   Ja   Ko   Nl   Pl   Pt   Region:us   Ro   Ru   Safetensors   Sharded   Tensorflow   Tr   Uk   Vi   Zh
Model Card on HF 🤗: https://huggingface.co/CohereForAI/aya-expanse-8b 

Aya Expanse 8B Benchmarks

LMSys ELO: 1180 vs 1272 (so35)-7.2%
MMLU Pro: 22.26
GPQA: 7.05
MUSR: 4.41
BBH: 28.52
IFEval: 63.59 vs 88 (so35)-27.7%
MATH Lvl 5: 8.61
LLME Score: 0.52098
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Aya Expanse 8B (CohereForAI/aya-expanse-8b)

Aya Expanse 8B Parameters and Internals

Model Type 
auto-regressive language model
Use Cases 
Areas:
Research, Commercial applications (non-commercial license required)
Applications:
Multilingual capabilities, Multilingual writing assistance, Multilingual question-answering
Primary Use Cases:
Multilingual language generation
Supported Languages 
en (English), fr (French), de (German), es (Spanish), it (Italian), pt (Portuguese), ja (Japanese), ko (Korean), zh (Chinese), ar (Arabic), el (Greek), fa (Persian), pl (Polish), id (Indonesian), cs (Czech), he (Hebrew), hi (Hindi), nl (Dutch), ro (Romanian), ru (Russian), tr (Turkish), uk (Ukrainian), vi (Vietnamese)
Training Details 
Data Sources:
Aya Evaluation Suite dataset
Methodology:
Auto-regressive, supervised finetuning, preference training, model merging
Context Length:
8000
Model Architecture:
Optimized transformer architecture
Input Output 
Input Format:
Models input text only
Accepted Modalities:
text
Output Format:
Models generate text only
LLM NameAya Expanse 8B
Repository 🤗https://huggingface.co/CohereForAI/aya-expanse-8b 
Model Size8b
Required VRAM16 GB
Updated2025-03-13
MaintainerCohereForAI
Model Typecohere
Model Files  4.9 GB: 1-of-4   4.9 GB: 2-of-4   5.0 GB: 3-of-4   1.2 GB: 4-of-4
Supported Languagesen fr de es it pt ja ko zh ar el fa pl id cs he hi nl ro ru tr uk vi
Gated ModelYes
Model ArchitectureCohereForCausalLM
Licenseproprietary
Context Length8192
Model Max Length8192
Transformers Version4.44.0
Tokenizer ClassCohereTokenizer
Padding Token<PAD>
Vocabulary Size256000
Torch Data Typefloat16

Best Alternatives to Aya Expanse 8B

Best Alternatives
Context / RAM
Downloads
Likes
Aya Expanse 8B8K / 16 GB800
Aya Expanse 8B Abliterated8K / 16 GB1355
Aya Expanse 8B Ungated8K / 16 GB111
Aya 23 8B Quantized8K / 8.1 GB133
...ng On New Data For Testing Aya8K / 18.1 GB80
Aya 23 8B 4bq8K / 5.7 GB140
AYA 8B8K / 32.1 GB140
Paya8K / 16 GB345
U3reb8K / 5.7 GB00
Aya 23 8B 4bit8K / 4.5 GB232

Rank the Aya Expanse 8B 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? 45005 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.
Release v20241227
  Support LLM Explorer