Unverified50% confidenceFactExact time
Nemotron-3.5-Super needs to read about 2.7GB of expert weights per token (22 experts × 40 layers × 3.10MB).
1
Sources
50%
Confidence
Long-term
Relevance
10/17/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe Nemotron 3 speaker diarization model has approximately 100 million parameters and requires only 4 GB of VRAM to run.69% similarUnverifiedThe 30B-class Nemotron series MoE models achieve at least 150 tokens per second on Mac with a memory footprint of approximately 18 GB.63% similarUnverifiedRelying on a compact MoE architecture with 12B activated parameters and optimized NVFP4 kernels in TensorRT-LLM, Nemotron generates 185 tokens/second on a single Blackwell B200 node, up to 3.2 times faster than dense competitors.63% similarUnverifiedAnatomy's SKILL.md is 31,600 bytes, plus about 300KB of nine reference guides loaded on demand at runtime; the injected prompt is about 7,700 input tokens, which takes 45 to 70 seconds to read on CPU, with generation speed of about 4 to 5.5 tokens per second.62% similarUnverifiedFor 240 million domain names, the memory footprint of a Succinct Trie can be compressed to only 2–4 bits per character, keeping the total within 1–2 GB.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1053036API
curl https://kongchang.com/api/v1/knowledge/claims/1053036MCP
get_claim(id=1053036)