Black Forest Labs' open-weights image generation model with frontier performance, available for non-commercial local deployment
Settled by community votes across 4 shared challenges, with an AI judge weighing in on each.
FLUX.2 [dev]
24.5
arena score
#17 of 44 in Text-to-Image
Skill signature
Not enough comparable category data
The chart appears once both models have ratings across at least three shared arena categories.
FLUX.2 [dev] Turbo
27.4
arena score
#4 of 44 in Text-to-Image
Vote tally
Where the votes landed
FLUX.2 [dev]
50.0%
win rate
Ties
0.0%
FLUX.2 [dev] Turbo
50.0%
win rate
50.0%
0.0% ties
50.0%
Shared challenges
4
Challenge by challenge
The strongest take from each model on every shared challenge, with the AI judge's read.
Geometric Composition
Text-to-Image“A glass cube on a wooden table. Inside the cube is a small blue sphere. On top of the cube sits a red book. A green plant is behind the cube, partially visible through the glass. Soft window light from the left.”
FLUX.2 [dev]
FLUX.2 [dev] Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Excellent soft lighting consistent with the prompt
- + Clean, minimalist aesthetic
- + Correct placement of all objects
- − The sphere appears to be levitating inside the cube rather than resting on the bottom
- − Glass cube has slightly rounded edges that look more like a container than a geometric cube
FLUX.2 [dev] Turbo
- + Highly realistic textures on the book and wooden table
- + Accurate glass refraction and dust details on the cube surface
- + Natural shadow and grounding for the blue sphere
- − The blue sphere looks slightly more like a marble than a simple sphere
- − The plant is slightly more obscured compared to Model A
Verdict: Both models followed the complex spatial prompt perfectly. FLUX.2 [dev] Turbo is the winner because of its superior texture work and realistic physics; specifically, the sphere sits naturally on the surface, whereas in FLUX.2 [dev] it appears to float. FLUX.2 [dev] Turbo also captured finer details like the wood grain and the embossed text on the book spine.
Candid Street Photography
Text-to-Image“A candid street photo of an elderly Japanese man repairing a red bicycle in light rain, reflections on wet pavement, shallow depth of field, 50mm lens, natural skin texture, imperfect framing, motion blur from passing cars, cinematic but realistic, no stylization.”
FLUX.2 [dev]
FLUX.2 [dev] Turbo
100% wins
0% ties
0% wins
AI Judge Analysis
FLUX.2 [dev]
- + Excellent motion blur on the passing cars communicating speed.
- + Highly realistic skin texture and facial lighting.
- + Natural, high-quality depth of field rendering.
- − The anatomy of the man's lower body and how he is interacting with the bike is physically confused.
- − The bicycle frame geometry is somewhat abstract and nonsensical in the center.
FLUX.2 [dev] Turbo
- + More coherent interaction between the man and the bicycle.
- + Great attention to detail with the tools and reflections on the pavement.
- + Captures the 'imperfect framing' request effectively by placing the subject slightly off-center.
- − The motion blur on the cars is much less pronounced, feeling more like a fast shutter speed than a 50mm candid.
- − The front wheel of the bicycle is clipping through the ground/pavement.
Verdict: Both models followed the prompt exceptionally well, capturing the atmosphere of a rainy Japanese street. FLUX.2 [dev] produced a more cinematically beautiful image with superior skin rendering and motion blur, but it failed on the structural logic of the bicycle. FLUX.2 [dev] Turbo provided a more grounded composition with better object logic, though it suffered from a clipping artifact where the front tire meets the road.
Fantasy Warrior
Text-to-Image“Close portrait of a battle-worn paladin in ornate engraved plate armor, hair braided with small beads, faint scars and dirt on the skin, warm torchlight reflecting off metal, shallow depth of field, bokeh sparks, lifelike eyes, highly detailed texture on leather straps and cloth underlayer.”
FLUX.2 [dev]
FLUX.2 [dev] Turbo
0% wins
0% ties
100% wins
AI Judge Analysis
FLUX.2 [dev]
- + Excellent warm lighting that authentically reflects off the plate armor.
- + Complex hair braiding with highly detailed bead work.
- + High contrast and rich textures on the leather and metal.
- − The scars look slightly painted on rather than recessed into the skin.
- − The bokeh sparks are a bit large and distracting compared to the subject.
FLUX.2 [dev] Turbo
- + Incredible skin texture and realistic depth to facial scars.
- + Highly detailed cloth and chainmail underlayers.
- + The colorful beads add a nice touch of personality to the character design.
- − The lighting is a bit cooler and flatter than Model A despite the torch presence.
- − Symmetry of the armor engraving feels slightly less organic.
Verdict: Both models captured the prompt with high accuracy, but FLUX.2 [dev] Turbo produced a more lifelike facial texture and better integration of scars into the skin. While FLUX.2 [dev] had superior warm lighting and reflections, the overall detail in the underlayers and the character realism of the Turbo model make it the stronger portrait.
Adorable Baby Animals in Sunny Meadow
Text-to-Image“Hyper-photorealistic scene of fluffy baby animals—a golden retriever puppy, tabby kitten, baby bunny, and red fox kit—with big expressive eyes and ultra-detailed soft fur, playfully chasing butterflies and tumbling together in a lush wildflower meadow, warm golden sunrise light with god rays and dew sparkles, joyful wholesome vibe, 8K masterpiece.”
FLUX.2 [dev]
FLUX.2 [dev] Turbo
AI Judge Analysis
FLUX.2 [dev]
- + Excellent fur detail and texture lighting.
- + Successfully included all four requested animal types.
- + Beautiful, clean composition with soft backlighting.
- − The animals are mostly sitting still rather than 'playfully chasing' and 'tumbling'.
- − Included a second redundant bunny instead of active movement.
FLUX.2 [dev] Turbo
- + Captures the 'playfully chasing' and 'tumbling' aspect of the prompt much better.
- + Dynamic poses and expressions convey a more joyful vibe.
- + Excellent rendering of dew sparkles and golden hour light.
- − The kitten's front paw is slightly distorted/blurry.
- − Slightly less 'hyper-realistic' in the fur textures compared to Model A.
Verdict: Both models followed the prompt well, but FLUX.2 [dev] Turbo better captured the requested action of the animals tumbling and chasing. While FLUX.2 [dev] produced a cleaner, more static portrait, FLUX.2 [dev] Turbo's dynamic composition feels more aligned with the 'sparkles' and 'joyful' tone requested.
Next steps
Explore each model
Distilled version of Black Forest Labs' FLUX.2 [dev] outperforming it at a cheaper price. Developed by fal.ai.