Developed in collaboration with Black Forest Labs, FLUX 3 Max delivers top-tier video quality, 22× faster generation than Seedance 2.5 at the same reported quality rating, and 11.5× faster inference than the original FLUX 3.
SAN FRANCISCO, Oct. 19, 2026 /PRNewswire/ -- fal, a leading generative media company, today announced FLUX 3 Max, a new video generation model delivering frontier visual quality at industry-leading inference speeds.
Post-trained by fal Research in collaboration with Black Forest Labs, FLUX 3 Max builds on the FLUX 3 foundation model through specialized post-training, reinforcement learning, and inference optimization. The collaboration brings together Black Forest Labs' expertise in foundation model development and fal's research in model optimization and high-performance inference.
In preliminary independent human preference evaluations on Artificial Analysis Video Arena, FLUX 3 Max achieved a 1,139 Elo rating across 4,868 votes, placing it among the highest-quality image-to-video models evaluated. In fal's latency benchmarks, the model generates a 10-second video in just 10.95 seconds, over 22× faster than Seedance 2.5, which achieves the same reported Elo rating.
Compared with the original FLUX 3, FLUX 3 Max also delivers significant improvements in both quality and speed:
- #1-tier image-to-video quality: Achieving 1,139 Elo on Artificial Analysis Video Arena in preliminary independent human preference evaluations, with overlapping confidence intervals among the highest-ranked models.
- 22× faster than Seedance 2.5: Generating a 10-second video in 10.95 seconds, compared with 277.80 seconds for Seedance 2.5, at the same reported image-to-video Elo rating.
- 50-point improvement in video quality: Increasing image-to-video Elo from 1,089 to 1,139 compared with the original FLUX 3.
- 11.5× faster inference than FLUX 3: Reducing generation latency from 126.44 seconds to 10.95 seconds.
These results demonstrate how specialized post-training and inference optimization can substantially improve both the quality and speed of frontier foundation models.
Combining Frontier Model Research with High-Performance Inference
FLUX 3 Max represents a collaboration between two research teams with complementary expertise: Black Forest Labs in foundation model development and fal Research in post-training, reinforcement learning, inference optimization, and systems engineering.
Rather than simply deploying an existing foundation model, fal Research has worked directly with FLUX 3 to improve visual fidelity, motion coherence, prompt adherence, and overall generation performance. The result is a model that delivers higher-quality video at significantly lower latency.
"We built FLUX 3 as a foundation for others to build on," said Robin Rombach, co-founder and CEO of Black Forest Labs. "FLUX 3 Max is a great example of what that looks like. fal took our model and customized it for the developers and enterprises on their platform. That's the kind of collaboration FLUX was designed for, and we're excited to see it reach more of the people building with it."
The collaboration reflects a broader approach to frontier AI development, where model capabilities continue to advance beyond initial foundation model training through specialized research and optimization.
"We believe some of the most important advances in generative AI will come from optimizing models and the systems that run them together," said Burkay Gur, CEO of fal. "Black Forest Labs has built an exceptional foundation model, and our research team has developed deep expertise in pushing the boundaries of quality and inference performance. FLUX 3 Max demonstrates what's possible when those capabilities come together: frontier video quality at speeds that fundamentally change how developers can build with generative AI."
Pushing the Frontier of Video Generation Quality and Speed
FLUX 3 Max builds on the research and engineering advances behind H3 Max, fal Research's previously released frontier video generation model.
Through specialized post-training and systems-level inference optimization, H3 Max demonstrated faster-than-real-time generation, producing video in just 3.49 seconds in fal's benchmark while achieving top-tier image-to-video quality in independent evaluations.
FLUX 3 Max extends this research approach to a different foundation model developed by Black Forest Labs, demonstrating how improvements in generation quality and inference performance can be achieved simultaneously.
Compared with the original FLUX 3, FLUX 3 Max improves image-to-video quality by 50 Elo points while delivering 11.5× faster generation. These improvements demonstrate the potential of targeted post-training and inference optimization to advance the speed–quality frontier for video diffusion transformers.
"Model performance doesn't stop improving when base model training ends," said Kaan Oktay, Head of Research at fal. "Through post-training and inference optimization, we're able to improve both the quality of generated video and the efficiency of the underlying model. With FLUX 3 Max, we've achieved a 50-point Elo improvement while making inference more than 11 times faster. And we're continuing to post-train the model to push performance even further."
This launch marks the first iteration of FLUX 3 Max. fal Research is continuing to post-train the model, with further improvements to visual quality, generation speed, and overall performance underway.
Together, fal and Black Forest Labs aim to make frontier video generation faster, more efficient, and more accessible to developers building the next generation of creative applications.
Availability
FLUX 3 Max is available beginning October 19, 2026, through fal's inference platform, giving developers access to frontier text-to-video, image-to-video, and keyframes-to-video generation through a unified API.
To celebrate the launch, FLUX 3 Max is available at 75% off for the first three weeks, making frontier-quality video generation more accessible to developers and creators.
Learn more and try FLUX 3 Max at fal.ai.
About fal
fal is a leading generative media platform powering AI applications across image, video, audio, 3D, and world models. Through its high-performance inference infrastructure and fal Research, the company enables developers and enterprises to build, deploy, and scale frontier generative AI applications.
Learn more at fal.ai.
About Black Forest Labs
Black Forest Labs is a frontier generative AI research company developing state-of-the-art models for visual content generation. Its FLUX family of models advances the capabilities of generative media, enabling developers, creators, and enterprises to build new visual experiences.
Learn more at bfl.ai.
SOURCE fal
Share this article
