CL-SN55 941 —
GPT-61S 885 —
GPT-56T 861 —
GPT-6S 825 —
CL-OP55X 818 -0.1%
MUSE-SPK 793 —
QWEN-38X 779 —
GPT-6A 778 —
CL-OP5H 711 -0.1%
CL-OP5X 709 -0.1%
KIMI-K3X 699 -0.1%
CL-OP46H 694 -0.1%
CL-FAB5H 692 -0.1%
CL-OP55H 692 -0.1%
CL-OP47H 683 -0.1%
GEM-37FH 648 -0.2%
GEM-38FH 633 -0.2%
INKL 531 —
CL-OP47 529 -0.2%
GPT-56S 529 -5.9%
GEM-3P 493 —
CL-OP46 488 —
CL-OP48 481 —
GEM-31P 453 -0.2%
GPT-6L 436 —
CL-SN55 941 —
GPT-61S 885 —
GPT-56T 861 —
GPT-6S 825 —
CL-OP55X 818 -0.1%
MUSE-SPK 793 —
QWEN-38X 779 —
GPT-6A 778 —
CL-OP5H 711 -0.1%
CL-OP5X 709 -0.1%
KIMI-K3X 699 -0.1%
CL-OP46H 694 -0.1%
CL-FAB5H 692 -0.1%
CL-OP55H 692 -0.1%
CL-OP47H 683 -0.1%
GEM-37FH 648 -0.2%
GEM-38FH 633 -0.2%
INKL 531 —
CL-OP47 529 -0.2%
GPT-56S 529 -5.9%
GEM-3P 493 —
CL-OP46 488 —
CL-OP48 481 —
GEM-31P 453 -0.2%
GPT-6L 436 —
← Back to feed

FLUX 3 Image Ships with Non-Destructive Multi-Step Editing and Bounding-Box Subject Control

Black Forest Labs has released FLUX 3 Image, the image-specific product built on its FLUX 3 foundation model. The release expands a family that already covers video and audio — FLUX 3 is a unified multimodal architecture trained across all three modalities using BFL’s Self-Flow approach.

The headline capability is non-destructive multi-step editing: users can apply successive text-guided edits and each pass modifies only the targeted region, leaving everything else unchanged. Prior image generation models collapsed the full scene on each edit, requiring users to either accept drift or preserve regions manually with masks. FLUX 3 Image makes the isolation automatic.

Subject Positioning via Bounding Boxes

The second major feature is compositional scene control. Users specify where each subject should appear using bounding boxes defined at generation time. BFL describes the result as full control over scene composition — a person in the foreground, an object in the upper-left, a background element constrained to a region — without post-generation repositioning.

The system supports text-to-image, image-to-image, text rendering, and photorealism in a single endpoint, covering the four primary commercial image generation use cases.

Output Resolution

FLUX 3 Image supports generation up to 4K resolution, with pricing tiered by resolution: approximately $0.041 for 768p, $0.048 for 1K, $0.10 for 2K, and $0.607 for 4K, with a 50% promotional launch discount. BFL published full credit and per-image pricing schedules at launch.

Open-Source Timeline

BFL has committed to releasing open-source weights for FLUX 3 Image within a few weeks of the commercial launch. In prior generations, select lightweight checkpoints — FLUX.1 Schnell and FLUX.2 Klein — carried Apache 2.0 licenses, while flagship tiers remained closed or non-commercial. BFL has not specified the license tier for the upcoming FLUX 3 Image open weights.

Context

BFL released FLUX 3, the underlying multimodal foundation, earlier in 2026 covering images, video, and audio in a unified model. Today’s release productizes the image surface specifically — naming it FLUX 3 Image to distinguish it from the full multimodal system. The model landed on Hacker News front page on October 2, 2026 with corroborating coverage from The Decoder, GIGAZINE, and kingy.ai.