Kapitsa Bolt

Nemotron 3.5 Bolt 74B-A9B (mixed 2/4/8-bit, 25.1 GB) running locally on Apple Silicon through Kapitsa. An agentic model in 24 GB.

kapitsa serve --model bolt

Speed

Mac Prefill (tok/s) Decode (tok/s)
M5 Max 128 GB 1,075 154
M5 Pro 48 GB 79
M4 Pro 48 GB 519 69
M3 Pro 36 GB 256 50

Architecture

Hybrid Mamba-2 MoE, 74B total parameters, 9B active per token, mixed 2/4/8-bit quantization.

Weights

nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for nace-ai/kapitsa-bolt