GPT Image 1.5vs.MOSS-Image-Base
Which one is better?
In our own test both are on par. MOSS-Image-Base has no arena rating yet. MOSS-Image-Base has no API price.
Test images
Same task, the first image of each. The task with the biggest difference is preselected.
Result for “Character and detail”:GPT Image 1.5 ahead
What the rating looks at: Anatomy, clothing, and spatial consistency.
“Ahead”: better in all three rating rounds. “Slightly ahead”: a higher value, but not in every round.
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
GPT Image 1.5: together, 1024x1024, n=1, seed=provider random, provider default steps, generated Aug 14, 2026, 12:17 PM. Source
MOSS-Image-Base: manual, local PyTorch 2.13.0, Diffusers 0.41.0; OpenMOSS-Team/MOSS-Image-Base@85975793d8b3d432f1b6e27f352abaa4c2a4c0d3; native 1024x1024, 1:1; one request, first output; steps=50, guidance=3.0, seed=921965437095743280; mps for text encoder and denoising; cpu for native VAE decode; bfloat16 for text encoder and transformer; float32 for VAE; unconditional skip layers=[22]; CPU operator fallback permitted, actual fallback use was not measured; official experimental MPS compatibility runtime OpenMOSS/MOSS-Image@1961ca3d111b828e65278656cde0c8f94484f4dd, generated Nov 2, 2026, 10:39 PM. Source
Protocol gradually-image-comparison-v1. Prompt SHA 256: 985a75b38642a9f0e5e0c0e0c67f1b7f67780d42bd2168fa24e9a7a7a94b0da6
GPT Image 1.5 image SHA 256: 0dff52ccd39400ac66ac8870c9ea937512ff1474688e70118688f420c38c2296
MOSS-Image-Base image SHA 256: 313d95f3e7766e861d46b0563c81bb2a02782708004c24c7aab0c46e8587ff83
Infographic
GPT Image 1.5: together, 1024x1024, n=1, seed=provider random, provider default steps, generated Aug 14, 2026, 12:17 PM. Source
MOSS-Image-Base: manual, local PyTorch 2.13.0, Diffusers 0.41.0; OpenMOSS-Team/MOSS-Image-Base@85975793d8b3d432f1b6e27f352abaa4c2a4c0d3; native 1024x1024, 1:1; one request, first output; steps=50, guidance=3.0, seed=2567672344276262662; mps for text encoder and denoising; cpu for native VAE decode; bfloat16 for text encoder and transformer; float32 for VAE; unconditional skip layers=[22]; CPU operator fallback permitted, actual fallback use was not measured; official experimental MPS compatibility runtime OpenMOSS/MOSS-Image@1961ca3d111b828e65278656cde0c8f94484f4dd, generated Nov 2, 2026, 10:43 PM. Source
Protocol gradually-image-comparison-v1. Prompt SHA 256: ed65cb4e6a51e2336f8b149ad4cbb16e020f07d527ad94362e719d2dcf91d94f
GPT Image 1.5 image SHA 256: 1c9ff59c32621c8b48b002dacb0ae66487262280c7c0a5b3d6456c81b248ebe9
MOSS-Image-Base image SHA 256: 3175582007cf377490ab51f73a57ebddb312f523ac5bfb54ae7598e72eb3edca
Product photography
GPT Image 1.5: together, 1024x1024, n=1, seed=provider random, provider default steps, generated Aug 14, 2026, 12:16 PM. Source
MOSS-Image-Base: manual, local PyTorch 2.13.0, Diffusers 0.41.0; OpenMOSS-Team/MOSS-Image-Base@85975793d8b3d432f1b6e27f352abaa4c2a4c0d3; native 1024x1024, 1:1; one request, first output; steps=50, guidance=3.0, seed=8912274556852986311; mps for text encoder and denoising; cpu for native VAE decode; bfloat16 for text encoder and transformer; float32 for VAE; unconditional skip layers=[22]; CPU operator fallback permitted, actual fallback use was not measured; official experimental MPS compatibility runtime OpenMOSS/MOSS-Image@1961ca3d111b828e65278656cde0c8f94484f4dd, generated Nov 2, 2026, 10:35 PM. Source
Protocol gradually-image-comparison-v1. Prompt SHA 256: c67fd71e91304458a2feef5c91efa8677260fab0c0099a29a2c52cb579f5b2f3
GPT Image 1.5 image SHA 256: 67cfa2f324447027ceb5f82e9446d40a1395b46ce425890a74606b4d97731cf6
MOSS-Image-Base image SHA 256: af9ea47eb1554bcaaaadf8860fecfa5bb69095a58bf346e3d9a73be8240a2a7c
Typography and layout
GPT Image 1.5: together, 1024x1024, n=1, seed=provider random, provider default steps, generated Aug 14, 2026, 1:59 PM. Source
MOSS-Image-Base: manual, local PyTorch 2.13.0, Diffusers 0.41.0; OpenMOSS-Team/MOSS-Image-Base@85975793d8b3d432f1b6e27f352abaa4c2a4c0d3; native 1024x1024, 1:1; one request, first output; steps=50, guidance=3.0, seed=9219943416610765538; mps for text encoder and denoising; cpu for native VAE decode; bfloat16 for text encoder and transformer; float32 for VAE; unconditional skip layers=[22]; CPU operator fallback permitted, actual fallback use was not measured; official experimental MPS compatibility runtime OpenMOSS/MOSS-Image@1961ca3d111b828e65278656cde0c8f94484f4dd, generated Nov 2, 2026, 10:31 PM. Source
Protocol gradually-image-comparison-v1. Prompt SHA 256: 2e012c3d14ee5dd02203a94f394ca80a5038cdebd6032b25a8b8a221f2255419
GPT Image 1.5 image SHA 256: 6a4cec86850fd371b65274198e2bf630b3cf36a096f9fb953d4bfe17c3222ebe
MOSS-Image-Base image SHA 256: e87b1d9d7aa09487c6b9be505746dacb2f3be551f29a5d85424a44a2a88089c5
The numbers at a glance
GPT Image 1.5
Rank 13 of 21
MOSS-Image-Base
not in the arena
GPT Image 1.5
Rank 29 of 108
MOSS-Image-Base
Rank 61 of 108
GPT Image 1.5
expensive
MOSS-Image-Base
no API price
Ranks among the 21 image models with an arena rating and the 108 with a GraduallyBench run; prices compared with 28 models with an API price.
GPT Image 1.5
- Subscription
- No OpenAI subscription includes this model. Through other providers from ≈ €6.36 a month (ElevenLabs Starter, with limited credits). 42 plans in Germany.
- Via API
- $0.13 per image, 100 images $13.30.
MOSS-Image-Base
- Subscription
- n/a No plan in Germany names exactly this model.
- Via API
- No API price known. Access: local.
- License
- Open weights.
Subscriptions only count plans that name exactly this model. Subscription prices include VAT, API prices are in dollars.
Arena rating (LMArena) from Nov 6, 2026 · GraduallyBench from Nov 3, 2026 · API prices from Nov 6, 2026 · Subscription database from Sep 29, 2026 (38 days old)
All data side by side
The better value is bold. For ratings only when the gap is clear.
| Kennzahl | GPT Image 1.5 | MOSS-Image-Base |
|---|---|---|
| Ratings | ||
| Arena ratingfrom votes by users, higher is better | 1,236 | n/a |
| Arena rank | Rank 13 of 21 | n/a |
| Arena votesmore votes, a safer rating | 173,798 | n/a |
| GraduallyBenchour own rating, 0 to 100 | 68 | 51 |
| GraduallyBench by task | ||
| Typography and layoutLegible text, hierarchy, and composition | 60 | 51 |
| Product photographyMaterials, reflections, and fine details | 52 | 36 |
| Character and detailAnatomy, clothing, and spatial consistency | 89 | 46 |
| InfographicNumbers, labels, and visual organization | 72 | 71 |
| API price | ||
| Per imagelower is better | $0.13 | n/a |
| 100 images | $13.30 | n/a |
| Profile | ||
| Released | December 2025 | Oct 16, 2026 |
| Weightsopen means: runs locally | proprietary | open |
| Lifecycle | Marked as deprecated by OpenAI; removed from the API on December 1, 2026 | no notice |
| Access | API, Web app | local |
| Parameters | not published | 5 billion |
| Largest output | 1,536 × 1,024 pixels | not consistently published |
| Plans in Germanynaming exactly this model | 42 | n/a |
API prices for one image at 1,024 × 1,024 pixels according to Artificial Analysis.
Sources: GPT Image 1.5 model source, status source (verified Aug 15, 2026) · MOSS-Image-Base model source (verified Nov 1, 2026)
Arena ratings by category
From blind votes on LMArena. “Ahead” means the lead is clear.
No shared arena score in the stored snapshot
The stored benchmark database contains no measurement for GPT Image 1.5 and MOSS-Image-Base from the same arena and category. Older model versions are not used as substitutes.
More comparisons
Compare GPT Image 1.5 and MOSS-Image-Base with the models people compare them with most often.