Cactus Compute

Needle 2: An Open 45M-Parameter Tool-Calling Model in a 14MB Binary

Cactus Compute has released Needle 2, an open 45M-parameter model for tool calling, device use, and structured extraction. The model ships as a 14MB binary, runs in about 28MB of RAM, and achieves high throughput on edge devices. It leads on Seal-Tools benchmarks but trails on BFCL v4.

Cactus Compute