Unverified50% confidenceFactExact time
1664x960 occupies 21.6GB in 24GB of VRAM (1.2GB baseline) and can only run when the GPU is truly idle; multiple rendering failures occurred because llama-swap and a resident engine slipped back onto the GPU mid-render.
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 a 16-bit precision 70B model on 2×80GB GPUs leaves only 2.9 GB of VRAM after the weights are loaded, accommodating only about 9,000 in-flight tokens, which in practice cannot serve requests effectively.73% similarVerifiedEven compressing the weights to 214GB using 4-bit quantization still far exceeds the capacity of a single 48GB GPU.72% similarUnverifiedTwo RTX PRO 6000 Blackwell cards each have 96GB of VRAM, 192 GB in total, but the 3-bit compressed model still requires 309.3 GiB and cannot be fully loaded into VRAM.71% similarUnverifiedPerforming LoRA fine-tuning on gpt-oss-20b with unquantized BF16 weights requires approximately 44 GB of VRAM.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1060352API
curl https://kongchang.com/api/v1/knowledge/claims/1060352MCP
get_claim(id=1060352)