Expired50% confidenceFactExact time
Flux 2 Dev introduces Mistral 7B as a text encoder, with approximately 7 billion parameters and a hidden layer dimension of 4096, causing VRAM consumption to rise sharply, potentially peaking at 24–32 GB or more.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/21/2026
First Seen
Valid until: 10/19/2026(expired)
Sources
Related Claims
UnverifiedFlux 引入 T5-XXL 作为文本编码器，代价是推理时显存需求增加约8-16GB69% similarUnverifiedFlux 基于扩散 Transformer（DiT）架构，与 SD 系列使用的 U-Net 架构不同，完整版需要约 24GB 显存，量化版可压至 8GB 左右63% similarUnverifiedFor 8 GB VRAM, Llama 3.1 8B Instruct, Qwen2.5 7B Instruct, and Mistral 7B with Q4_K_M quantization are recommended.63% similarUnverifiedThe combined unified memory of two DGX Sparks totals approximately 256 GB.61% similarUnverified传统Stable Diffusion使用CLIP-L/14作为文本编码器，嵌入维度768，最大token长度77，参数量约1.2亿61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/583461API
curl https://kongchang.com/api/v1/knowledge/claims/583461MCP
get_claim(id=583461)