aifollow.news Search
Back IT之家 科技新闻
IT之家 科技新闻· · Original publication time

Black Forest Labs releases open-weight FLUX 3 Image Klein image modelMachine translation

Automatically verified and published · Generated and evidence-checked automatically; not reviewed by a human.

AI-assisted summary

FLUX 3 Image Klein has 5.6B parameters and supports text-to-image generation, image editing and multi-reference generation. Its INT8 version can run through ComfyUI or Diffusers on an NVIDIA GeForce RTX GPU with 8GB of VRAM; in the reported test, it generated a 1MP image in under 9 seconds on an RTX 4070 using ComfyUI.

Article · AI translation

IT Home reported on October 21 that visual AI lab Black Forest Labs released its FLUX 3 Image Klein image generation model with open weights on the 15th of this month, local time.

FLUX 3 Image Klein has 5.6B parameters and is optimized for consumer GPUs. Its BF16 weights require 12.2GB of memory, while its FP8 weights require 5.6B of memory. The INT8 version of the model can run through ComfyUI or Diffusers on an NVIDIA GeForce RTX GPU with just 8GB of VRAM.

According to tests, the streamlined INT8 version takes about 4s to generate a 1MP image on a GeForce RTX 5090, and less than 9s to generate a 1MP image on a GeForce RTX 4070 using ComfyUI.

FLUX 3 Image Klein supports text-to-image generation, image editing, and generation from multiple references. It natively supports 4K resolution and can deliver pixel-precise editing and consistent pattern placement within specified bounds. Black Forest Labs says it outperforms larger open-source models in multiple scenarios and offers a better balance of speed and quality.

References

https://bfl.ai/blog/flux-3-image-klein

Advertising notice: External links in this article (including but not limited to hyperlinks, QR codes, and passcodes) are provided to convey more information and save readers time in making selections. The results are for reference only. All IT Home articles containing external links include this notice.

Found an error? Send a correction