Waifmark 2

Building the Future of Local Companions

The First Agentic + Roleplay Benchmark for 100% Local Small (V)LLMs.

View Leaderboard → Control Center Repo

Top 5 right now

—

Loading leaderboard…

Run it

↓ Download benchmark suite

DownloadGrab Waifmark as a .zip (or press button above!)
Unzip then open a terminal inside benchmark/.
RunPaste the commands below, and put your API keys in .env.
Openthe Control Center at 127.0.0.1:8001 to run the vllm/llama.cpp server, model and benchmark.

TERMINALCopy

cd benchmark
python3 -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env  # add OPENROUTER_API_KEY / HF_TOKEN
waifmark              # ← boots at http://127.0.0.1:8001

GitHub

qubixal/waifmarkCode, scoring methodology, raw project code
IssuesBug reports and model requests
leaderboard.jsonRaw published scores (see View Leaderboard for actual leaderboard)