Unverified50% confidenceBenchmarkExact time
When generating about 1 MP images on an RTX PRO 6000, Qwen-Image-2.1's average peak allocated VRAM was 36.801 GiB, a bit over thirty percent of the 95.0 GiB of available VRAM reported by PyTorch.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/16/2026
First Seen
Valid until: 10/30/2026
Sources
Qwen-Image-2.1 Tested on RTX PRO 6000 with 15 Images
rss10/16/2026
Related Entities
Related Claims
UnverifiedOn an RTX PRO 6000, using BF16 precision, 40 steps and a batch size of 1, Qwen-Image-2.1 took an average of 9.378 seconds and a median of 9.375 seconds to generate 14 images of about 1 MP.73% similarUnverified在RTX 3080实测中，用户采用0.2百万像素/608×352分辨率的生成设置68% similarUnverifiedThe Qwen3.8-27B checkpoint occupies 28.75 GiB on disk, and after loading in vLLM the weights plus non-torch overhead are about 29.34 GiB, leaving 53.8 GiB on a 96 GB RTX Pro 6000 for the KV cache, equivalent to about 767,317 tokens.68% similarUnverified一张1920×1080的PNG图片动辄几百KB，而原生XML描述同样内容只需几KB68% similarUnverifiedImage models can maintain INT8 or FP8 precision when offloaded, and Z-Image generation time on a B200 with a 10 GiB VRAM limit dropped from 38.4 seconds to 2.3 seconds67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1042282API
curl https://kongchang.com/api/v1/knowledge/claims/1042282MCP
get_claim(id=1042282)