What runs on your machine

31 of the 38 models BenchClub runs fit in 22 GB

$ uv tool install "git+https://github.com/holaboss-ai/holaOS@feat/benchclub-cli#subdirectory=packages/benchclub"
ModelMemory neededArena · Creative writing
Qwen3.8-27BQwen22 GB1377
Qwen3.5-27BQwen22 GB1362
gemma-3-27b-itGoogle22 GB1345
gemma-3-12b-itGoogle12 GB1331
Mistral-Small-3.2-24B-Instruct-2506Mistral AI19 GB1304
gemma-3n-E4B-itGoogle13 GB1287
gemma-3-4b-itGoogle8 GB1270
granite-4.2-8bIBM Granite8 GB1235
granite-4.2-3bIBM Granite8 GB1204
gpt-oss-20bOpenAI18 GB1202
phi-4Microsoft13 GB1181
Llama-3.1-8B-InstructMeta Llama8 GB1154
Llama-3.2-3B-InstructMeta Llama8 GB1094
Mistral-7B-Instruct-v0.2Mistral AI8 GB1068
Llama-3.2-1B-InstructMeta Llama8 GB1033
Qwen3-0.6BQwen8 GB—
Qwen3-14BQwen13 GB—
Qwen2.5-7B-InstructQwen8 GB—
Qwen3-8BQwen8 GB—
Qwen3.5-2BQwen8 GB—
Qwen3-4BQwen8 GB—
Qwen3-1.7BQwen8 GB—
Qwen3-4B-Instruct-2507Qwen8 GB—
NVIDIA-Nemotron-3-Nano-4B-BF16NVIDIA8 GB—
Qwen2.5-Coder-14B-InstructQwen13 GB—
Qwen2.5-Coder-7B-InstructQwen8 GB—
MiniCPM5-2BOpenBMB8 GB—
DeepSeek-R1-Distill-Qwen-1.5BDeepSeek8 GB—
Phi-3.5-mini-instructMicrosoft8 GB—
gemma-3-1b-itGoogle8 GB—
LFM3-1.7BLiquid AI8 GB—
Compare the top 3

Needs more than 22 GB