Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.benchgen.com/llms.txt

Use this file to discover all available pages before exploring further.

Upload a Model

Eval supports two ways to bring in a model: connecting an external API endpoint or uploading model weights directly.

Option 1: Connect an External API

Use this for hosted models (OpenAI, Anthropic, your own inference endpoint, or a fine-tuned model from Train).
  1. Go to Eval → Models → Add Model.
  2. Select API connection.
  3. Enter your endpoint URL and authentication token.
  4. Click Test connection to verify.
  5. Save — the model is now available for benchmark runs.

Option 2: Upload Model Weights

Use this for local checkpoints or merged LoRA adapters exported from Train.
  1. Go to Eval → Models → Add Model.
  2. Select Upload weights.
  3. Upload your model files (supported formats: safetensors, GGUF).
  4. Select the base architecture.
  5. BenchGen spins up a temporary inference endpoint automatically.

Next Steps