Overview
FLUX.2 [max] is the flagship image generation and editing model from Black Forest Labs, offering the highest fidelity in the FLUX.2 series. It is built for complex, multi-reference workflows, allowing creators to maintain strict character and object consistency across different environments and lighting conditions. Featuring real-time web grounding and deep prompt comprehension, it provides maximum precision for professional production, serving as a heavier, more capable alternative to [Flux.2 pro].
Specifications · Text + image to image
- Input mode
- Text + image to image
- Accepts
- reference image (up to 8)
- Aspect ratios
- 16:9, 9:16, 4:3, 3:4, 1:1
Pricing
Build with this model: the Flux.2 [max] API on the Hedra Developer Platform.
Flux.2 [max] API →Real output · generated with Flux.2 [max]
Best of Flux.2 [max]
Prompting
Prompt tips
- Use JSON for granular control: Instead of relying purely on natural language, use structured JSON parameters to explicitly dictate lighting setups, material properties, and spatial reasoning.
- Stack your references: Take advantage of the multi-reference support to lock in a character's face with one image while dictating the scene's style with another.
- Trigger live context: Explicitly ask the model to "integrate current weather conditions" or reference real-time data to activate its grounded generation feature.
- Be structurally explicit: Write long, detailed prompts defining exact camera angles, layout patterns, and typography placement—the model has a massive context window and will follow strict instructions.
About the model
Questions, answered
What is Flux.2 [max] best used for?
Flux.2 [max] is Black Forest Labs' highest-fidelity image model, designed for commercial-grade production. It excels at complex multi-reference editing, accurate typography rendering, and photorealistic detail. The community relies on its ability to maintain character and style consistency across large batches, making it suitable for hero images, marketing assets, and final production renders where visual quality is the primary requirement.
When was Flux.2 [max] released, and what is its lineage?
Black Forest Labs launched the FLUX.2 family on November 25, 2025, and released the flagship FLUX.2 [max] on December 16, 2025. It succeeds the FLUX.1 generation, which includes Flux 1.1 Pro and Flux Kontext Max. It sits at the top of its current family, above [Flux.2 pro] and [Flux.2 flex], and is frequently compared by the community to top-tier alternatives like Nano Banana Pro.
How should I format prompts for Flux.2 [max]?
Flux.2 [max] supports JSON-formatted structured prompts, which give you granular control over complex compositions, lighting, and materials. You can also input up to 10 reference images simultaneously to maintain character consistency or apply style transfers across generations. For strict brand guidelines, the model natively supports exact HEX color codes (like #02eb3c) directly in the prompt text. See the official prompting guidelines for detailed examples.
Can Flux.2 [max] generate images based on current events?
Yes. A distinct feature of Flux.2 [max] is Grounded Generation. The model can search the web for real-time context, allowing you to generate accurate visualizations of current events or up-to-date real-world information. This bypasses the typical training-data cutoff limitations found in most other image models.
