- Home
- Image model comparison
- FLUX.2 Klein 4B vs. Kandinsky-5.0-T2I-Lite-SFT
Black Forest LabsKandinsky LabImage model matchup
FLUX.2 Klein 4Bvs.Kandinsky-5.0-T2I-Lite-SFT
Which one is better?
In our own benchmark GraduallyBench Image v2, FLUX.2 Klein 4B is slightly ahead, but not certainly. Neither is in the arena. Neither has an API price.
FLUX.2 Klein 4B
Rank 61 of 108
Kandinsky-5.0-T2I-Lite-SFT
Rank 82 of 108
Ranks among the 21 image models with an arena rating and the 108 with a GraduallyBench Image v2 run; prices compared with 28 models with an API price.
Test images from GraduallyBench Image v2
Same task, the first image of each. The task with the biggest difference is preselected.


Result for “Typography and layout”:FLUX.2 Klein 4B ahead
What the rating looks at: Legible text, hierarchy, and composition.
“Ahead”: the 95 % ranges of both values do not overlap. “Slightly ahead”: a higher value, but the ranges overlap.
Test protocol of the images
Each test attempt uses the same archived prompt and one request. The first output of that attempt counts. Repeats after technical test failures are disclosed for the affected images. Files are normalized to 1,024 × 1,024 pixels without cropping. Provider resolution, randomness, quality tier and prompt processing can still differ. The images are a practical visual test, not a reconstruction of the arena configuration.
Character and detail
FLUX.2 Klein 4B: fal, requested 1024x1024, n=1, seed=707400387, Fal's published 4B text-to-image endpoint; default four inference steps, estimated cost $0.005000, generated Nov 1, 2026, 9:01 PM. Source
Kandinsky-5.0-T2I-Lite-SFT: manual, local PyTorch 2.13.0, Diffusers 0.36.0; kandinskylab/Kandinsky-5.0-T2I-Lite-sft-Diffusers@c0158b5b20858d1fc02088db7ed4d805a296498b; native 1024x1024, 1:1; one request, first output; steps=50, guidance=3.5, seed=89463658088346984; mps for text encoder and denoising; cpu for native VAE decode; bfloat16 for text encoders; float32 for transformer and CPU VAE; native tokenizer truncation: CLIP 80 to 77 tokens; archived input unchanged, generated Nov 2, 2026, 8:12 PM. Source
Protocol gradually-image-comparison-v1. Prompt SHA 256: 985a75b38642a9f0e5e0c0e0c67f1b7f67780d42bd2168fa24e9a7a7a94b0da6
FLUX.2 Klein 4B image SHA 256: 170cebb7cf75bc033a92371199812674f9cfa9641d5e9fb13947ccf892d4ed53
Kandinsky-5.0-T2I-Lite-SFT image SHA 256: 8d193e702ad1aa0fccdb45f5d1474d60e21d455fa58d13470edd2796af0f2aed
Infographic
FLUX.2 Klein 4B: fal, requested 1024x1024, n=1, seed=290140178, Fal's published 4B text-to-image endpoint; default four inference steps, estimated cost $0.005000, generated Nov 1, 2026, 9:01 PM. Source
Kandinsky-5.0-T2I-Lite-SFT: manual, local PyTorch 2.13.0, Diffusers 0.36.0; kandinskylab/Kandinsky-5.0-T2I-Lite-sft-Diffusers@c0158b5b20858d1fc02088db7ed4d805a296498b; native 1024x1024, 1:1; one request, first output; steps=50, guidance=3.5, seed=649868597087646690; mps for text encoder and denoising; cpu for native VAE decode; bfloat16 for text encoders; float32 for transformer and CPU VAE, generated Nov 2, 2026, 8:21 PM. Source
Protocol gradually-image-comparison-v1. Prompt SHA 256: ed65cb4e6a51e2336f8b149ad4cbb16e020f07d527ad94362e719d2dcf91d94f
FLUX.2 Klein 4B image SHA 256: 42b729a5a72207b5a92464f3ecf1713459ab1b105c3041c14b75c7622630015e
Kandinsky-5.0-T2I-Lite-SFT image SHA 256: e102016732db9e8e463504a83ac0770ff8a3c4c77918143d0f9dae8498c79101
Product photography
FLUX.2 Klein 4B: fal, requested 1024x1024, n=1, seed=1254275381, Fal's published 4B text-to-image endpoint; default four inference steps, estimated cost $0.005000, generated Nov 1, 2026, 9:00 PM. Source
Kandinsky-5.0-T2I-Lite-SFT: manual, local PyTorch 2.13.0, Diffusers 0.36.0; kandinskylab/Kandinsky-5.0-T2I-Lite-sft-Diffusers@c0158b5b20858d1fc02088db7ed4d805a296498b; native 1024x1024, 1:1; one request, first output; steps=50, guidance=3.5, seed=4636200764941689006; mps for text encoder and denoising; cpu for native VAE decode; bfloat16 for text encoders; float32 for transformer and CPU VAE; native tokenizer truncation: CLIP 80 to 77 tokens; archived input unchanged, generated Nov 2, 2026, 8:02 PM. Source
Protocol gradually-image-comparison-v1. Prompt SHA 256: c67fd71e91304458a2feef5c91efa8677260fab0c0099a29a2c52cb579f5b2f3
FLUX.2 Klein 4B image SHA 256: 8fae4a7f6abb7467281c439dc93aa656031626f88c8a54f95f49850cbf045580
Kandinsky-5.0-T2I-Lite-SFT image SHA 256: 73b31dc5a48059f7712c6a125ac4855f2401133101172c4bfbaa11b76c5ceee1
Typography and layout
FLUX.2 Klein 4B: fal, requested 1024x1024, n=1, seed=1270698692, Fal's published 4B text-to-image endpoint; default four inference steps, estimated cost $0.005000, generated Nov 1, 2026, 9:00 PM. Source
Kandinsky-5.0-T2I-Lite-SFT: manual, local PyTorch 2.13.0, Diffusers 0.36.0; kandinskylab/Kandinsky-5.0-T2I-Lite-sft-Diffusers@c0158b5b20858d1fc02088db7ed4d805a296498b; native 1024x1024, 1:1; one request, first output; steps=50, guidance=3.5, seed=378697477899808767; mps for text encoder and denoising; cpu for native VAE decode; bfloat16 for text encoders; float32 for transformer and CPU VAE; three local execution attempts: first aborted on an MPS dtype error; second returned an unsaved latent but the runner read the wrong output field; accepted first saved image after correcting that field with unchanged prompt, seed and FP32 sampling settings; lost latent content remains unknown, generated Nov 2, 2026, 7:52 PM. Source
Protocol gradually-image-comparison-v1. Prompt SHA 256: 2e012c3d14ee5dd02203a94f394ca80a5038cdebd6032b25a8b8a221f2255419
FLUX.2 Klein 4B image SHA 256: 7c2c3e30c20c98e90033087b2b95b151bddc3252dd4aed7092fcbca4ae4ea96e
Kandinsky-5.0-T2I-Lite-SFT image SHA 256: a84280dd219692f7b20722a37491d6788d1ecdc594f816fa1203981642501ab2
Subscription and API
FLUX.2 Klein 4B
- Subscription
- No Black Forest Labs subscription includes this model. Through other providers from $4.99 a month (Poe Basic, with limited points). 12 plans in Germany.
- Via API
- No API price known (access: API, local).
- Weights
- Open weights.
Kandinsky-5.0-T2I-Lite-SFT
- Subscription
- n/a No plan in Germany names exactly this model.
- Via API
- No API price known (access: local).
- Weights
- Open weights.
Subscriptions only count plans that name exactly this model. Subscription prices include VAT, API prices are in dollars.
GraduallyBench Image v2 from Nov 7, 2026 · Subscription database from Sep 29, 2026 (49 days old)
All data side by side
The better value is bold. For ratings only when the gap is clear.
| Kennzahl | FLUX.2 Klein 4B | Kandinsky-5.0-T2I-Lite-SFT |
|---|---|---|
| Ratings | ||
| Arena ratingfrom votes by users, higher is better | n/a | n/a |
| Arena rank | n/a | n/a |
| Arena votesmore votes, a safer rating | n/a | n/a |
| GraduallyBench Image v2our own benchmark, 0 to 100 | 46 | 27 |
| GraduallyBench Image v2 by task | ||
| Typography and layoutLegible text, hierarchy, and composition | 55 | 18 |
| Product photographyMaterials, reflections, and fine details | 59 | 28 |
| Character and detailAnatomy, clothing, and spatial consistency | 48 | 37 |
| InfographicNumbers, labels, and visual organization | 21 | 27 |
| API price | ||
| Per imagelower is better | n/a | n/a |
| 100 images | n/a | n/a |
| Profile | ||
| Released | n/a | n/a |
| Weightsopen means: runs locally | open | open |
| Access | API, local | local |
| Parameters | 4 billion | 6 billion |
| Largest output | not consistently published | not consistently published |
| Plans in Germanynaming exactly this model | 12 | n/a |
API prices for one image at 1,024 × 1,024 pixels according to Artificial Analysis.
Sources: FLUX.2 Klein 4B model source (verified Nov 1, 2026) · Kandinsky-5.0-T2I-Lite-SFT model source (verified Nov 1, 2026)
Arena ratings by category
From blind votes on Artificial Analysis and LMArena; the sources are named under the card. “Ahead” means the lead is clear.
No shared arena score in the stored snapshot
The stored benchmark database contains no measurement for FLUX.2 Klein 4B and Kandinsky-5.0-T2I-Lite-SFT from the same arena and category. Older model versions are not used as substitutes.
More comparisons
Compare FLUX.2 Klein 4B and Kandinsky-5.0-T2I-Lite-SFT with the models people compare them with most often.