Unverified50% confidenceEventExact time
Over the past 72 hours, Meta, Nvidia, Reflection and DeepSeek have successively released open-source large models, with parameter counts ranging from 74B to 816B.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/18/2026
First Seen
Valid until: 11/1/2026
Sources
Related Entities
Related Claims
UnverifiedRunning an open-source model at the 70-billion-parameter scale requires at least approximately 140GB of VRAM in FP16 precision, i.e., 8 H100s in parallel.64% similarUnverifiedThe DeepSeek distilled versions range in parameter count from 1.5B to 70B.64% similarUnverified部署DeepSeek-R1 671B MoE模型需要至少数张高端GPU（如H100或A100），显存需求超过320GB62% similarUnverifiedMainstream GPU inference solutions (such as deployments based on NVIDIA H100) typically achieve only a few dozen to one or two hundred tokens per second on 27B-scale models.62% similarUnverifiedAirLLM 能够在单张仅有4GB显存的GPU上完成70B大模型的推理61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1064203API
curl https://kongchang.com/api/v1/knowledge/claims/1064203MCP
get_claim(id=1064203)