FLUX 3 Image Generation: What Is Announced and When It Opens
Stylized still-image collage representing FLUX 3 image generation
FLUX 3 is not only a video model. Black Forest Labs also announced FLUX 3 Image for synthesis across styles, aspect ratios, and resolutions, with stronger complex-prompt handling than earlier FLUX versions. The catch: image early access was not released with the first video gate. BFL said it would follow in the weeks after the July 23, 2026 video announcement.
Note: Until image early access opens, there are no broad public hands-on results. TheFluxTrain support for FLUX 3 is coming soon.
Quick answer: FLUX 3 Image is announced for generation (and editing) with varied styles, aspect ratios, resolutions, and improved complex prompts, including multilingual text rendering. Video early access opened first; image early access was expected in the following weeks. Public pricing and TheFluxTrain integration are not live yet. See the FLUX 3 model page.
Pillar: FLUX 3 capabilities and release status. Related: image editing · text rendering.
What can FLUX 3 Image generate?
From the announcement:
- Varied styles
- Multiple aspect ratios and resolutions
- Better handling of complex prompts than earlier FLUX versions
- Accurate text in multiple languages (claimed)
Exact max resolution, seed behavior, and safety filters are not fully documented for general users.
How will image generation connect to video?
The multimodal pitch is a shared foundation across stills and motion. A practical creator loop looks like:
- Generate or edit a still with FLUX 3 Image
- Animate it with FLUX 3 image-to-video
- Chain more clips under the 20-second rule
That loop only works end-to-end when both image and video access are in your hands.
What should you prepare before image early access?
- Prompt library for your brand or film look
- Aspect ratio list (1:1, 4:5, 9:16, 16:9, print sizes)
- Reference boards for characters and products
- Notes on where readable type must appear (posters, packaging)
If typography is the job, read the text rendering guide first.
How does this compare to earlier FLUX models?
Earlier FLUX models (Dev, Schnell, and the ones many of us trained LoRAs on) were still-image focused and already strong for a lot of creators. FLUX 3 Image is positioned as the next synthesis tier inside a multimodal family that also does video and audio. Do not assume drop-in parity with your old LoRA stacks until weights, APIs, and fine-tuning paths are published. FLUX 3 Dev open weights are planned with no date.
Availability checklist
| Piece | Status (July 2026 announcement window) |
|---|---|
| FLUX 3 Video | Gated early access |
| FLUX 3 Image | Early access expected after video |
| Public pricing | Unannounced |
| FLUX 3 Dev open weights | Planned, no date |
| On TheFluxTrain | Coming soon |
Frequently asked questions
Is FLUX 3 Image available now?
Video early access opened first. Image early access was announced for the following weeks. Check BFL and the model page for current status.
Can FLUX 3 Image do editing too?
Yes, editing is part of the announced image capability. See FLUX 3 image editing.
Does it render text in images?
BFL claims accurate multilingual text rendering. Verify on real outputs when you have access.
Will my old FLUX LoRAs work?
No compatibility statement has been published for FLUX 3.
Is FLUX 3 Image open source?
Open-weight FLUX 3 Dev is planned later. “Open-weight” is the accurate term until a license is published.
Can I use FLUX 3 Image on TheFluxTrain today?
No. Status stays on the FLUX 3 model page.
What resolutions are supported?
Varied resolutions are announced; a public max-resolution table is not.
Where should I start?
Read the release overview, then this page for stills-focused planning.