@mijanlab/llmtest
v1.0.5
Published
Fast zero-dependency CLI for benchmarking OpenAI-compatible LLM endpoints
Maintainers
Readme
LLMtest — Node.js Beta
This beta ports the CLI runtime to Node.js so Python, pip, virtualenvs, and PyInstaller are not required.
Requirements
- Node.js 18+
- npm
One-line install (recommended)
Checks for Node.js 18+, installs it automatically if missing, then installs llmtest globally via npm.
macOS / Linux:
curl -fsSL https://raw.githubusercontent.com/mijanlab/LLMtest/main/install.sh | bashWindows (PowerShell):
irm https://raw.githubusercontent.com/mijanlab/LLMtest/main/install.ps1 | iexThen run:
llmtestOr:
llmtest https://api.openai.com/v1 YOUR_API_KEYManual npm installation
If you already have Node.js 18+ installed:
npm i -g @mijanlab/llmtest@latestInstall this ZIP locally (offline / pre-publish testing)
unzip LLMtest-beta.zip
cd LLMtest-beta
npm install -g .Included beta features
- OpenAI-compatible
/modelsdiscovery - Streaming
/chat/completionsbenchmark - TTFT, total latency, and tokens/sec measurement
- Concurrent model testing
- Model filtering and run count
- JSON, CSV, Markdown, and HTML reports
- Interactive setup wizard
llmtest --updatellmtest --uninstallllmtestandllm-testexecutable aliases
Beta note
The current Python implementation on main is not modified by this ZIP. Validate provider compatibility and report parity before replacing the production branch.
