Flux 1.1 Pro: The New Benchmark in AI Image Generation

Meet Flux 1.1 Pro: The Best AI Image Generator You Can't Run - Decrypt

Black Forest Labs, the studio behind the Fluxfamily of AI image generators, announced last week the release of Flux 1.1 [Pro]. This comes just two months after the release of its original family of models including Flux 1 Pro (a closed source model with industry-leading capabilities), Flux 1 Dev (a noncommercial, open source model) and Flux Schnell (a fully open source model). The Flux models marked a major leap in generative AI technology with their text generation capabilities, prompt adherence and overall image quality. Even the smaller models, Flux Dev and Flux Schnell, generated results on par with generations from MidJourney, and way better than the outputs provided by SD3, Stability's much anticipated evolution over SDXL, which turned out to be somewhat underwhelming. The new model has already made a mark, securing the top Elo score in the Artificial Analysis image arena -- a leading benchmarking platform for AI models. It has outperformed every other text-to-image model on the market while being almost as fast as its smallest model. The graph below shows the Elo score (image quality) on the Y axis and the generation speeds on the X axis. MidJourney enthusiasts may notice that their model is not represented -- it's so slow it's literally off the chart. However, its Elo Score sits somewhere around 1100 points, just below Ideogram V2. The new Flux Pro stands out in terms of pricing, with Flux1.1 Pro costing $0.04 per image -- lower than many other models on the market, including the original Flux1 Pro. This pricing structure makes it a strong competitor over other paid services like MidJourney and Ideogram, which cost $96 and $84 a year each. The MidJourney and Ideogram options are also slower and offer a higher cost per token. Sadly, Flux1.1 Pro cannot be run locally. Unlike its less powerful open source counterparts, such as the FLUX1 [Dev] and FLUX1 [Schnell] models, this new pro version is a closed source model, limiting users to access it through platforms like Together AI, Replicate, Fal AI, and Freepik. It cannot be fine tuned or personalized. For those considering trying the model, some of these platforms offer a few credits for free generations, but once those are depleted, the best service according to our own criteria is Freepik. That's because its Mystic workflow dramatically enhances generations with higher details and better aesthetics. There are no announcements regarding an open source 1.1 version of the FLUX1 [Dev] or FLUX1 [Schnell] models, but it's clear that Black Forest Labs is focusing its efforts on great models for image and video creators. We tried the new Flux model and the results were satisfactory. It is not a generational leap -- like the move from SDXL from Flux -- but it is certainly a welcomed upgrade. It is overall very realistic, has great text generation capabilities and is very creative in artistic tasks and styles. It is a good, versatile model that provides fast generations without compromising quality. Realism Prompt: "Polaroid photo with VSCO filter, 1990, woman, night, flash photo, blonde, young face, beautiful shadows, tropical plants, inside an apartment, DSLR, camera flash, holding a handwritten sign on a notebook saying 'Verification for Decrypt October 7, 2024.' The woman is doing the peace sign with her other hand." The model excels at producing realistic images, improving upon the airbrushed look of the initial Flux models. While not perfect, the results are highly convincing, especially with proper prompting. At first glance, these images -- both generated with Flux 1.1 Pro -- could pass for real without nitpicking small details. The lettering is consistent with the prompt, and hand rendering has improved, though it's not quite perfect. It's important to note that these are not hand-picked samples but the first two generations. When working with generative AI, the best results typically come after several generations and edits. The lighting is consistent with a camera flash, focusing on the subject without illuminating the whole room. The VSCO filter enhances realism, and prompt adherence is excellent. Comparing Flux 1.1 to Flux 1 shows that the generations are quite similar in terms of realism at first glance. However, using the same prompt, the new model produces a more natural pose and a more consistent body. For example, Flux 1 generated what could appear to be an additional leg, which Flux 1.1 avoided. This improvement has more to do with accuracy than overall realistic aesthetic. Prompt Adherence Prompt: "A white cat playing the piano, wearing sunglasses and a hat, wearing purple Hawaiian style, full body shot against a gray studio background with lighting elements and a pterodactyl hanging from the ceiling, commercial video screengrab. The wall has the text 'Emerge by Decrypt'" Flux 1.1 takes prompt adherence a step further compared to Flux 1 Pro, successfully incorporating more elements into the scene without missing the mark. Our first prompt for Flux 1 didn't include the lighting elements or the pterodactyl. Additionally, the new generation is more realistic and feels more natural. Spatial Awareness Prompt: "A dog standing on top of a TV showing the word 'Decrypt' on the screen. On the left there is a woman in a business suit holding a coin, on the right there is a robot standing on top of a first aid box. The overall scenery is surreal." In terms of spatial awareness, Flux 1.1 and Flux 1 are comparable. Both generated all elements without issues. However, Flux 1.1 Pro seems superior when considering additional details. For instance, there's less prompt spilling (when the model takes elements from the prompt and uses them in other areas). In the Flux 1.1 generation, the woman holds one coin with no visible additional coins, while Flux 1 generated a stash of coins next to the dog. Moreover, the error with the additional hand in Flux 1 Pro isn't present in the newer model, and the surreal style is better represented in the Flux 1.1 generation. Conclusion Flux 1.1 Pro is overall more consistent and logical in its generations. If you can't run a local model, it's a very good competitor. It understands natural language, making it suitable for beginners, though this isn't its primary strength. MidJourney tends to be more creative while enhancing poor prompts. However, Flux 1.1 Pro is cheaper, faster, and generally better in quality than any current model, potentially making it the best option for those seeking good prompt adherence, quality, and text generation capabilities. For those willing to pay for the model, any of the current options does the job. We liked the service provided by Fal.AI because it provides more control than the others. However, Freepik seems to be the best option for those who want a more pro experience. While slightly more expensive, it's significantly more versatile, offering not only image generation services but also additional features like image upscaling, outpainting, draft-to-image generations, a background remover, and a library of content for experimentation.

Move over, Midjourney -- Flux just got a huge upgrade for generating AI images

Flux, the AI generation model from Black Forest Labs, is one of the best AI image generators, particularly because it's open source and so is widely available. Now, there is a new version -- and it has taken things to a new level. Flux 1.1 Pro offers improved realism and significantly faster image creation than previous versions of Flux Pro. In my tests, I found the pro model as fast as generating lower-resolution images with Schnell, the smallest of the Flux models. Initially, Freepik will be the only generation platform with access to Flux 1.1 Pro although it will also be available through API calls from other providers. To mark the launch Freepik is making the new model accessible to free users until the weekend. According to Black Forest Labs, Flux 1.1 Pro provides six times faster generation than its predecessor while also improving image quality, prompt adherence and diversity. Previously dubbed '"Blueberry mode" in user ranking tests, where people have to vote between two images made using the same prompt, this new model outperforms the previous generation of Flux 1 Pro, Ideogram 2 and Midjourney 6.1 -- topping the leaderboard. I was given early access to Flux 1.1 Pro and during my testing I found the new model was much better at following a descriptive prompt, reducing the need to use "AI enhancement" for prompts in the Freepik platform. The images generated also looked more natural but there were some issues around generating faces if multiple people were in the frame. The better prompt adherence also leads to a more literal outcome, so if you want to capture a smartphone-style photograph, for example, you are better off specifying the phone model and camera type than just saying 'smartphone photo of...' In one example I generated an image of a group of friends taking a selfie with a smartphone and the first version had several phones with people capturing their own image within the frame, rather than one person taking an image of everyone. I did find that the speed improvements led to a drop in the impressive levels of photorealism compared to Flux 1 Pro, but that might just be a case of needing some fine-tuning. Black Forest Labs makes new models widely available. The exception is the Pro versions, which are largely restricted to AI platforms due to the hardware requirements to run the model. This should still be the case with Flux 1.1 Pro, but you will likely find it on many platforms. Freepik will have Flux 1.1 Pro as part of its AI image service. It will be free initially but after Friday night it will only be available to paying subscribers. You will likely also find Flux 1.1 Pro on API-like platforms such as FAL and eventually on services like Poe, HuggingChat and NightCafe.

Black Forest Labs releases Flux 1.1 Pro and an API

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Black Forest Labs (BFL), a startup founded by the creators of the popular Stable Diffusion AI image generation model that underpins many AI image generation apps and services (such as Midjourney), has announced the release of a new, faster text-to-image model called Flux 1.1 Pro, and with it, a paid application programming interface (API) on which developers can build third-party apps powered by the model (or incorporate it into their existing apps). This means that a company that offers creative tools can add Flux as an option to their offerings, if they (and by extension, their end users) are willing to pay the API costs. Individual users can access the new Flux 1.1 Pro model not through Black Forest Labs's site, but rather, through partners together.ai, Replicate, fal.ai, and Freepik. Some of these services refer to the model under a different name, such as "Flux Fast." No details were immediately provided about Flux 1.1 Pro's training dataset, an issue of contention for generative AI companies with the original Stability AI and rival Midjourney being sued by artists who accuse the firms and others of violating their copyright by scraping and training en masse without consent or compensation on human-created images posted to the web. One key class action lawsuit against Stability AI and Midjourney remains in court. The news comes following the success of Flux's initial open source text-to-image AI model which powers Elon Musk's Grok 2 chatbot from xAI and available to subscribers of his social network X. Unlike its earlier model Flux.1, which was open source and free for anyone to download, fine-tune, customize, and otherwise use for all commercial or personal uses as they saw fit, the new Flux 1.1 Pro model appears to be, like Flux 1.0 Pro, a paid proprietary offering only. However, it is still available for commercial and enterprise usage. BFL sees the launch of its API and Flux 1.1 Pro as major steps in its growth as a company, offering both developers and enterprises access to powerful and customizable tools for image generation. Codenamed "Blueberry," Flux 1.1 Pro takes the new top spot on the Artificial Analysis image arena leaderboard Flux 1.1 Pro improves on the earlier Flux 1.0 Pro model by delivering six times faster generation speeds, while also enhancing image quality, prompt adherence, and diversity. It enables workflows that prioritize speed without sacrificing quality, generating output three times faster than its predecessor. Additionally, BFL announced an update for the original Flux 1.0 Pro, doubling its generation speed to improve efficiency across the board. The performance of Flux 1.1 Pro has been validated through its secret debut on Artificial Analysis, an independent third party benchmark platform for comparing AI model performance, where the model was tested in the days prior to today's announcement under the code name "blueberry." (Some erroneously speculated on X that this was OpenAI testing Sora following its tests of the o1 LLM as "strawberry.") As of October 1, 2024, Flux 1.1 Pro holds the highest ELO score on the platform at 1153, surpassing other generative models in terms of visual fidelity and prompt accuracy, including Midjourney 6.1 (ELO score of 1100) and Ideogram v2 (score of 1108). The ELO third-party benchmark was established earlier this summer of 2024 by Artificial Analysis co-founder and CEO Micah Hill-Smith and co-founder and Product Lead George Cameron, and uses human ratings of pairs of images to derive its scores. For users demanding high-resolution outputs, Flux 1.1 Pro will soon support ultra-high-resolution images (up to 2k), maintaining its precision and speed through upcoming API updates. BFL API offers developers AI image generation starting at 4 cents per image Complementing the Flux 1.1 Pro release is the BFL API in beta, which brings BFL's generative capabilities directly to businesses and developers looking to integrate state-of-the-art image generation into their own applications. The API offers advanced customization, enabling users to adjust model choice, resolution, and content moderation to meet their specific needs. It also promises scalability, making it suitable for projects ranging from small-scale to enterprise-level. BFL's API comes with competitive pricing, making it attractive for users seeking high-quality outputs without excessive costs. For example, the Flux 1.1 Pro image generation is priced at USD $0.04 per image, while the older Flux 1.0 Pro is available at $0.05 per image. Developers can begin integrating the API today, and BFL promises ongoing improvements as the beta progresses. The company envisions its API opening the door to countless creative applications, especially in industries like design, advertising, and entertainment, where demand for high-quality AI-generated media continues to grow. Building on initial strong success Black Forest Labs is no stranger to the spotlight. Just two months earlier, the company secured $31 million in seed funding, led by Andreessen Horowitz (a16z), with backing from high-profile investors such as Brendan Iribe, Michael Ovitz, and Garry Tan. As reported by VentureBeat, the launch of BFL and its earlier Flux 1.0 model was widely seen as a milestone in the AI community. BFL co-founders Robin Rombach, Patrick Esser, and Andreas Blattmann brought their expertise from Stability AI, the team behind Stable Diffusion, into this new venture, with a vision for more accessible, open-source generative AI tools. Flux 1.0, which came in three variants (Flux 1.0 Pro, Flux 1.0 Dev, and Flux 1.0 Schnell), gained early praise for its 12-billion parameter architecture and its ability to match or even surpass the output quality of competing models like MidJourney and DALL-E. The open-source nature of these models, especially Flux 1.0 Dev and Flux 1.0 Schnell, positioned BFL as a critical player in the debate over open-source versus proprietary AI. Industry context and competition Black Forest Labs' move to launch Flux 1.1 Pro comes at a time of heightened competition in the generative AI media space, with many creators looking to harness text-to-image AI models alongside image-to-video models such as those from Pika, Runway, and Luma. Midjourney and Ideogram are both competing directly with Flux in the paid proprietary text-to-image AI model space, while Stability AI continues to offer both open source and proprietary models under the leadership of former Weta (film special effects) CEO Prem Akkaraju and Hollywood director James Cameron (Titanic, Avatar, Terminator), who recently joined the company's board. This integration into a social platform signals how generative AI is becoming more accessible to mainstream users, raising the stakes for other players in the field. What's next for BFL? Looking ahead, Black Forest Labs is already working on expanding its generative AI capabilities beyond images. The company has set its sights on text-to-video systems, a development that could further solidify its leadership in the AI-driven media space. If successful, BFL's expansion into video could further disrupt industries such as advertising, content creation, and virtual reality. It also comes as Midjourney is reportedly pursuing generative AI video models and hardware as well. For now, Flux 1.1 Pro and the BFL API represent significant advancements in generative technology, offering users faster, more efficient tools without compromising quality. Whether through their own API or partner platforms like together.ai, Replicate, fal.ai, and Freepik, BFL is looking to make Flux 1.1 Pro the AI image generation model of choice for most users. As BFL continues to push the boundaries of generative AI, the company is also expanding its workforce, seeking talented innovators to join its mission. Interested candidates can explore open positions via the company's website.

FLUX 1.1 Pro vs FLUX LORA vs Midjourney V6.1 vs Mystic V2 Performance Comparison

If you're curious about how today's leading AI image generators compare -- FLUX 1.1 Pro, FLUX LORA, Midjourney V6.1, and Mystic V2 -- CyberJungle's new video has you covered. This in-depth comparison reveals how each model brings its unique strengths to the art of image creation, offering insights into where they shine and what sets them apart. FLUX 1.1 Pro, fondly called the "blueberry" model, stands out with its speed and efficiency, while its sibling, FLUX LORA, offers nuanced capabilities for diverse creative needs. Midjourney V6.1 continues to impress with its artistic sensitivity and interpretation, making it a favorite for creative projects. Mystic V2 completes the lineup with an emphasis on photorealism, pushing the boundaries of what's possible in hyper-realistic imagery. When it comes to prompt understanding and execution, each model demonstrates distinct characteristics: In the realm of photorealism, Mystic V2 takes the lead, producing outputs that closely mimic real-world imagery. FLUX 1.1 follows closely behind, delivering impressive results that blur the line between artificial and authentic. Interestingly, Midjourney's latest version seems to have taken a step back in terms of photorealism compared to its predecessors, indicating potential areas for future improvement. Detail accuracy becomes crucial when generating complex scenes. The FLUX models demonstrate a strong ability to maintain accuracy across intricate details. In contrast, both Midjourney and Mystic V2 face some challenges in this area, particularly when dealing with highly complex scenarios. For users focused on illustrative art, Midjourney V6.1 emerges as a clear frontrunner. Its ability to create stunning artistic styles sets it apart from the competition. The FLUX and Mystic models, while capable in many areas, show room for growth in this particular aspect of image generation. Here are a selection of other articles from our extensive library of content you may find of interest on the subject of AI image generators : Each model comes with its own set of advantages and limitations: The AI image generation field continues to evolve rapidly. FLUX is reportedly working on a video model, potentially expanding its capabilities beyond static images. Midjourney's upcoming version 7 promises to introduce a storytelling mode, which could transform how users interact with the platform. Mystic V2, while powerful in many respects, would benefit from workflow improvements and more consistent character feature generation to enhance its overall user experience. Selecting the ideal AI model for your needs requires careful consideration of your specific requirements. There is no one-size-fits-all solution, as each model offers unique strengths that cater to different user needs. As these models continue to develop, user feedback will play a crucial role in shaping their future iterations. Staying informed about the latest capabilities and updates of these AI models will empower you to make the best choice for your creative or professional projects. The key lies in aligning the model's strengths with your specific needs, ensuring that you harness the full potential of AI-driven image generation in your work.