Needle 2: Open Tool-Calling Model Runs in 28MB RAM
TL;DR. Needle 2, a 45M-parameter open tool-calling model, now ships as a 14MB binary and operates within 28MB of RAM. - This compact AI model can execute a full session on resource-constrained devices. - Its design focuses on efficient tool-calling capabilities in a small footprint. - The model's minimal resource requirements enable broader edge AI applications.
- Needle 2 is a 45M-parameter open tool-calling model.
- It ships as a 14MB binary.
- The model runs a full session in just 28MB of RAM.
- Its efficiency makes it suitable for on-device and edge deployments.