Unverified50% confidenceFactExact time
Nemotron 3.5 Super VL's backbone was initialized from Nemotron 3 Super Base, distilled 1T tokens via Nemotron 3 Ultra Base, extended to a 1M context, then underwent 67B tokens of software engineering mid-training, and only then was vision training done.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedNVIDIA is reportedly training Nemotron 4, while its Nemotron 3.5 has already seen widespread usage.79% similarUnverifiedNemotron系列基于Transformer架构构建，在训练阶段利用DGX SuperPOD等超大规模GPU集群，并通过NeMo Framework进行全流程管理75% similarUnverifiedNemotron 3.5 Super VL uses a mixture-of-experts (MoE) architecture, routing each token to only about one-tenth of the expert networks.75% similarUnverifiedNVIDIA used knowledge distillation to compress the capabilities of the Nemotron 3 family (including Ultra) into Nemotron 3.5 Super VL.73% similarUnverifiedNemotron 3.5 Super VL is specifically optimized for Agent workflows, emphasizing low latency and high token throughput.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1049460API
curl https://kongchang.com/api/v1/knowledge/claims/1049460MCP
get_claim(id=1049460)