v2.2 · 2026-06-12
Aya 23 8B
aya-23-8b
Aya 23 8B is a 8B text chat model profile for local AI planning. This page records provider, license, context, quantization, hardware-fit estimates, setup hints, source links, and caveats so readers do not choose a model by name alone.
PROVIDER
Cohere Labs
FAMILY
Aya
MODEL TYPE
text-chat
PARAMETERS
8B
MODALITIES
text
ARCHITECTURE
Auto-regressive optimized transformer instruction model trained for multilingual generation across 23 languages.
CONTEXT WINDOW
8,192 tokens
TRAINING TOKENS
—
RELEASE DATE
2024-05-23
LICENSE & LINKS
CC-BY-NC-4.0 with Cohere Labs Acceptable Use Policylicense text
MODEL CARD
HUGGING FACE
GGUF REPOSITORIES
bartowski/aya-23-8B-GGUF community conversion
Community GGUF conversion with Q4_K_M, Q5_K_M and Q8_0 file listings; original HF model access is gated by Cohere Labs terms.
SETUP HINTS
OLLAMA
Run `ollama run aya:8b-23`.LM STUDIO
Use a community Aya 23 8B GGUF such as the bartowski conversion and select Q4_K_M or Q5_K_M.
LLAMA.CPP
Use a standard text GGUF loader against `bartowski/aya-23-8B-GGUF`, subject to the original model license terms.QUANTIZATION FILE SIZES (GGUF)
Q4_K_M
5.05 GB
Q5_K_M
5.8 GB
Q8_0
8.54 GB
Quantization note: Q4_K_M for local multilingual chat and summarization; Q5_K_M if memory allows.
RAM / VRAM ESTIMATES
MIN RAM
10 GB
COMFORTABLE RAM
16 GB
MIN VRAM
6 GB
COMFORTABLE VRAM
12 GB
These are conservative local-inference estimates, not official hardware requirements.
HARDWARE FIT
CPU only (no GPU)
yes
8 GB RAM
limited
16 GB RAM
yes
32 GB RAM
yes
8 GB VRAM
limited
12 GB VRAM
yes
24 GB VRAM
yes
Apple Silicon (unified memory)
yes
Hardware fit values are conservative local-inference estimates based on GGUF size plus runtime and KV-cache overhead. Actual requirements depend on context length, quantization, runtime, GPU offload, and other running apps.
BEST FOR
·Noncommercial multilingual local chat
·Translation-style assistance across the source-listed 23-language set
·Multilingual summarization
·Multilingual evaluation and research
·Language coverage demos
AVOID IF
·You need commercial use rights
·You cannot accept gated HF model terms
·You need tool calling
·You need vision
·You need the newer Aya Expanse or Aya Vision family
CAVEATS
·Chosen over Yi 1.5 9B Chat for the multilingual slot because Aya 23 8B verifies 23 supported languages; Yi 1.5 is stronger as an English/Chinese bilingual permissive model but is less broad for this bucket.
·Aya 23 8B is gated on Hugging Face and governed by CC-BY-NC-4.0 plus Cohere Labs Acceptable Use Policy, so it is not suitable for commercial use.
·The cited GGUF is a community conversion.
·The model card identifies Aya 23 as an older version relative to Aya Expanse/Aya Vision, but Aya 23 8B remains a clear local text-only multilingual target.
·RAM and VRAM estimates are derived from GGUF/Ollama package size plus conservative overhead; KV cache, context length, batch size, runtime and GPU offload choices can materially change memory needs.
SOURCE URLS
FIELD EVIDENCE
canonicalModelCardUrlhuggingface.co/CohereLabs/aya-23-8B
architecturehuggingface.co/CohereLabs/aya-23-8B
parameterSizeBhuggingface.co/CohereLabs/aya-23-8B
releaseDatearxiv.org/abs/2405.15032
contextWindowTokenshuggingface.co/CohereLabs/aya-23-8B
languageshuggingface.co/CohereLabs/aya-23-8B
q4FileSizeGbhuggingface.co/bartowski/aya-23-8B-GGUF
q5FileSizeGbhuggingface.co/bartowski/aya-23-8B-GGUF
q8FileSizeGbhuggingface.co/bartowski/aya-23-8B-GGUF
LANGUAGES
ArabicChinese simplified and traditionalCzechDutchEnglishFrenchGermanGreekHebrewHindiIndonesianItalianJapaneseKoreanPersianPolishPortugueseRomanianRussianSpanishTurkishUkrainianVietnamese
CAPABILITIES
No special capabilities flagged.