Expired50% confidenceFactExact time
Full fine-tuning of Llama-3 (8B) requires approximately 80 GB or more of VRAM, while Llama-3 (70B) requires more than 400 GB.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/15/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/13/2026(expired)
Sources
Related Claims
UnverifiedLlama 3.1 7B量化后仅需约4-5GB显存即可运行81% similarUnverifiedAfter 4-bit quantization, Llama 3.3 70B has a model size of approximately 40 GB and can run on a Mac Studio with 64 GB of memory or a single professional GPU with 48 GB of VRAM.79% similarVerifiedThe weights alone of the Llama-3 70B model require approximately 140 GB of VRAM at FP16 precision.77% similarUnverifiedMeta Llama 3的405B参数版本以BF16精度加载需约810GB内存，量化至INT4精度需约200GB73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/514350API
curl https://kongchang.com/api/v1/knowledge/claims/514350MCP
get_claim(id=514350)