aya-expanse-8b
Cohere Labs- Arena score
- 1185
- Among open models
- #115of 148
- Among all on Arena
- #334
- Votes
- 9,818
Measured on Aya Expanse 8B, the served version, not a local build. LMArena, CC BY 4.0, published Nov 11, 2026.
Score by category
CategoryScoreOpen rankArena rankVotes
- Legal & government1225#106#309551
- Science1204#109#3221.7K
- Medicine & health1203#103#303470
- Russian1197#108#3141.5K
- Long prompts1191#111#3261.5K
- English1191#117#3395.5K
- German1187#85#266240
- Writing & literature1185#111#3272.7K
- Chinese1181#109#321618
- Non-English1180#111#3234.3K
- Software & IT1177#115#3322.5K
- Applied math1177#113#3261.1K
- Business & finance1176#110#3221.1K
- Math1168#112#3251.3K
- Creative writing1168#112#3281.5K
- Coding1160#118#3411.6K
- Multi-turn1160#115#3341.7K
- Hard prompts1156#117#3392.6K
- Instructions1155#117#3364K
- Hard prompts, English1154#121#3441.6K
- Expert1152#108#312600
- Media & sports1150#112#3311.8K
- Japanese1121#74#249231
- Excluding ties1053#115#3346.3K
About as strong, smaller
Compare
granite-4.2-3bIBM Granite12923.7B
Built on it
Ways to run it
- GGUF · llama.cpp, Ollama, LM Studiobartowski/aya-expanse-8b-GGUF41 builds
- MLX · Apple silicon: mlx-lm, LM Studiomlx-community/aya-expanse-8b-4bit3 builds
- AWQ · vLLM, SGLangOrion-zhen/aya-expanse-8b-AWQ2 builds
- GPTQ · vLLM, SGLangkevinbazira/aya-expanse-8b-gptq-4bit1 build
Most downloaded
BenchClub runs
Not runnable yet
Runs open once a aya-expanse-8b build is registered.
More from Cohere Labs
See more
tiny-aya-global31.7K3.3B

aya-expanse-32b3.6K32.3B

c4ai-command-r-plus202104B
Other 3–9B models
See more
NVIDIA-Nemotron-3-Nano-4B-BF16NVIDIA1.9M4.0B
granite-4.1-8bIBM Granite79.4K8.8B
Molmo2-8BAi247.6K8.7B
Spark-X2.6-4BXHToken32.9K4.1B