Head-to-head comparison

GPT Image 2 vs. Stable Diffusion 3.5 Large

Compare shared capability tests, arena scores with documented retrieval dates, task-specific image pricing, and verified specifications. Research benchmarks, preference arenas, and our n = 1 samples remain separate.

shared result series: 13
shared sample prompts: 4
latest arena retrieval: 2026-09-22

First image modelSecond image model

Specifications and cost

General specifications and task-specific pricing

A model can have one price for generating new images and another for editing. Prices therefore remain tied to the corresponding arena configuration.

Specifications and pricing for GPT Image 2 and Stable Diffusion 3.5 Large
Attribute	GPT Image 2	Stable Diffusion 3.5 Large
Provider	OpenAI	Stability AI
Released	April 2026	October 22, 2024
Access	API, Web app	API, local
Lifecycle	No separate lifecycle notice	No separate lifecycle notice
Weights	Proprietary	Open weights
Parameters	Not published	8.1 billion
Maximum output	4K	1 Megapixel
Arena configuration	High	No separate configuration named
Text-to-image price	$211 per 1,000 images	$65 per 1,000 images
Editing price	$211 per 1,000 images	No arena entry
Technical source	Model sourceVerified 2026-08-15	Model sourceVerified 2026-08-15

GPT Image 2Stable Diffusion 3.5 Large

Text to Image ArenaUSD per 1,000 images at 1,024 × 1,024 pixels

GPT Image 2$211

Stable Diffusion 3.5 Large$65

Key to the marks:better valuebetter, but not conclusivetie

Prompt adherence and specialist tests

Shared capability benchmarks

Both checkpoints completed the same four Gradually sample tasks. For each task, 30 archived first outputs are ranked in three differently ordered, model-blind passes. The published 0-100 value normalizes the mean rank. It must still be read as an n = 1 sample per model. No exact shared result from the integrated research benchmarks is available for this pair.

GPT Image 2Stable Diffusion 3.5 Large

Gradually sample test: Typography and layoutMean of three blind ranks across all 30 first outputs, n = 1 each

GPT Image 241.4 / 100

Stable Diffusion 3.5 Large2.3 / 100

Gradually sample test: Product photographyMean of three blind ranks across all 30 first outputs, n = 1 each

GPT Image 283.9 / 100

Stable Diffusion 3.5 Large0 / 100

Gradually sample test: Character and detailMean of three blind ranks across all 30 first outputs, n = 1 each

GPT Image 293.1 / 100

Stable Diffusion 3.5 Large26.4 / 100

Gradually sample test: InfographicMean of three blind ranks across all 30 first outputs, n = 1 each

GPT Image 278.2 / 100

Stable Diffusion 3.5 Large4.6 / 100

Key to the marks:better valuebetter, but not conclusivetie

Capability benchmark scores for GPT Image 2 and Stable Diffusion 3.5 Large
Benchmark	GPT Image 2	Stable Diffusion 3.5 Large
Gradually sample test: Typography and layout	Higher blind-rank score41.4 / 100 Mean rank: 18 (20 / 24 / 10) Five rubric scores Headline: 1.67 / 2 Subline: 1.67 / 2 Text control: 1.67 / 2 Visual brief: 1.33 / 2 Layout: 1.33 / 2	2.3 / 100 Mean rank: 29.33 (30 / 28 / 30) Five rubric scores Headline: 0.67 / 2 Subline: 0.67 / 2 Text control: 0.67 / 2 Visual brief: 1.33 / 2 Layout: 1 / 2
Gradually sample test: Product photography	Higher blind-rank score83.9 / 100 Mean rank: 5.67 (11 / 2 / 4) Five rubric scores Subject: 1.67 / 2 Time: 1.67 / 2 Watch details: 1.67 / 2 Photography: 1.67 / 2 Exclusions: 1.67 / 2	0 / 100 Mean rank: 30 (30 / 30 / 30) Five rubric scores Subject: 1 / 2 Time: 1.67 / 2 Watch details: 1.67 / 2 Photography: 1.67 / 2 Exclusions: 1 / 2
Gradually sample test: Character and detail	Higher blind-rank score93.1 / 100 Mean rank: 3 (3 / 2 / 4) Five rubric scores Scene: 1 / 2 Wardrobe: 1 / 2 Anatomy: 1 / 2 Props: 1 / 2 Environment: 1 / 2	26.4 / 100 Mean rank: 22.33 (13 / 24 / 30) Five rubric scores Scene: 1 / 2 Wardrobe: 1 / 2 Anatomy: 1 / 2 Props: 1 / 2 Environment: 1 / 2
Gradually sample test: Infographic	Higher blind-rank score78.2 / 100 Mean rank: 7.33 (7 / 9 / 6) Five rubric scores Title: 1.67 / 2 Sequence: 1.67 / 2 Labels: 1.67 / 2 Diagram: 1 / 2 Clarity: 1.67 / 2	4.6 / 100 Mean rank: 28.67 (29 / 28 / 29) Five rubric scores Title: 0.67 / 2 Sequence: 0.33 / 2 Labels: 0.33 / 2 Diagram: 1 / 2 Clarity: 0.33 / 2
Key to the marks:better valuebetter, but not conclusivetie

Independent preference tests

Shared arena benchmarks

Artificial Analysis and Arena derive scores from blind comparisons. Raters see outputs for the same prompt and choose the image they prefer. Separate series cover text rendering, photorealism, portraits, art, commercial design, and image editing. Confidence intervals indicate how certain a measured lead is.

9 shared result series

GPT Image 2Stable Diffusion 3.5 Large

Text to Image ArenaBlind preference from the same prompt, higher is better

GPT Image 21,171 Elo

Stable Diffusion 3.5 Large839 Elo

LMArena Text to Image: 3D modelingBlind preference from the same prompt, higher is better

GPT Image 21,361.089

Stable Diffusion 3.5 Large963.888

LMArena Text to Image: ArtBlind preference from the same prompt, higher is better

GPT Image 21,365.89

Stable Diffusion 3.5 Large997.656

LMArena Text to Image: CartoonBlind preference from the same prompt, higher is better

GPT Image 21,399.599

Stable Diffusion 3.5 Large945.681

LMArena Text to Image: Commercial designBlind preference from the same prompt, higher is better

GPT Image 21,389.09

Stable Diffusion 3.5 Large937.143

LMArena Text to Image: OverallBlind preference from the same prompt, higher is better

GPT Image 21,381.149

Stable Diffusion 3.5 Large938.344

LMArena Text to Image: PhotorealismBlind preference from the same prompt, higher is better

GPT Image 21,380.609

Stable Diffusion 3.5 Large928.163

LMArena Text to Image: PortraitsBlind preference from the same prompt, higher is better

GPT Image 21,427.407

Stable Diffusion 3.5 Large941.209

LMArena Text to Image: Text renderingBlind preference from the same prompt, higher is better

GPT Image 21,426.635

Stable Diffusion 3.5 Large921.527

Key to the marks:better valuebetter, but not conclusivetie

Arena benchmark scores for GPT Image 2 and Stable Diffusion 3.5 Large
Benchmark	GPT Image 2	Stable Diffusion 3.5 Large	Source
Text to Image Arena	Higher score1,171 Elo Rank 3, 95% confidence interval: ±9, 15,176 ratings	839 Elo Rank 124, 95% confidence interval: ±9, 5,130 ratings	Artificial AnalysisRetrieved 2026-09-22
LMArena Text to Image: 3D modeling	Higher score1,361.089 Rank 3, 95% confidence interval: ±9.3, 7,425 ratings	963.888 Rank 76, 95% confidence interval: ±15.92, 1,629 ratings	LMArenaRetrieved 2026-09-19
LMArena Text to Image: Art	Higher score1,365.89 Rank 2, 95% confidence interval: ±7.91, 10,214 ratings	997.656 Rank 73, 95% confidence interval: ±9.48, 5,340 ratings	LMArenaRetrieved 2026-09-19
LMArena Text to Image: Cartoon	Higher score1,399.599 Rank 3, 95% confidence interval: ±6.48, 31,995 ratings	945.681 Rank 76, 95% confidence interval: ±6.35, 12,274 ratings	LMArenaRetrieved 2026-09-19
LMArena Text to Image: Commercial design	Higher score1,389.09 Rank 2, 95% confidence interval: ±6.32, 30,140 ratings	937.143 Rank 77, 95% confidence interval: ±11.07, 3,051 ratings	LMArenaRetrieved 2026-09-19
LMArena Text to Image: Overall	Higher score1,381.149 Rank 3, 95% confidence interval: ±4.38, 78,731 ratings	938.344 Rank 77, 95% confidence interval: ±4.52, 23,396 ratings	LMArenaRetrieved 2026-09-19
LMArena Text to Image: Photorealism	Higher score1,380.609 Rank 3, 95% confidence interval: ±6.29, 30,097 ratings	928.163 Rank 77, 95% confidence interval: ±7.09, 9,405 ratings	LMArenaRetrieved 2026-09-19
LMArena Text to Image: Portraits	Higher score1,427.407 Rank 3, 95% confidence interval: ±8.27, 15,130 ratings	941.209 Rank 76, 95% confidence interval: ±8.91, 5,704 ratings	LMArenaRetrieved 2026-09-19
LMArena Text to Image: Text rendering	Higher score1,426.635 Rank 3, 95% confidence interval: ±6.8, 29,289 ratings	921.527 Rank 77, 95% confidence interval: ±10.86, 3,222 ratings	LMArenaRetrieved 2026-09-19
Key to the marks:better valuebetter, but not conclusivetie

Visual comparison

Sample images

Our images complement preference data, but do not replace it. The prompt, request count, and selection rule match. Provider resolution, quality tier, randomness, and prompt processing remain visible limitations, so this is not a fully controlled laboratory test.

Largely standardized comparison protocol. Both models receive the same archived prompt, exactly one request, and the first result counts. Files are normalized to 1,024 × 1,024 pixels without cropping. Provider resolution, randomness, quality tier, and prompt processing can still differ. These images are a practical visual test, not a reconstruction of the arena configuration.

Character and detail

Anatomy, clothing, and spatial consistency

Prompt and settings

Create a square cinematic full-body portrait of a fictional bicycle courier waiting under a transparent umbrella at a rainy tram stop in Hamburg at blue hour. She wears a mustard raincoat, navy trousers, red sneakers, a silver helmet, and carries a teal messenger bag. Keep both hands visible, render the bicycle correctly, use realistic wet-street reflections, and include no readable brand names.

GPT Image 2. Provider Together, 1024x1024, n=1, seed=provider random, provider default steps, model ID openai/gpt-image-2, generated Aug 14, 2026, 12:13 PM. Source
Stable Diffusion 3.5 Large. Provider fal.ai, requested 1024x1024, n=1, seed=7087247023770270000, estimated cost $0.068157, model ID fal-ai/stable-diffusion-v35-large, generated Aug 14, 2026, 2:37 PM. Source

Protocol gradually-image-comparison-v1, selection rule: first result.

Prompt SHA 256: 985a75b38642a9f0e5e0c0e0c67f1b7f67780d42bd2168fa24e9a7a7a94b0da6

GPT Image 2, image SHA 256: 82f6e2369ab4fb6efb0ef48feeff0d14f1d6087c64c35113a983456f0f0879ac

GPT Image 2, request SHA 256: db4ca97dc43effd578430b77f8ca559c7dd5cbe60e663ba9ecc2e89a33480b0d

Stable Diffusion 3.5 Large, image SHA 256: e748006fe56e442ad347e3f994f18b86ee1739f20dcee063aa7a03ee710c5d54

Stable Diffusion 3.5 Large, request SHA 256: 5d005ee24926d68b7d2299b8ecbe4e14919f652761f6de2568a043f2a6258da0

Infographic

Numbers, labels, and visual organization

Prompt and settings

Create a square German infographic titled SO FUNKTIONIERT PHOTOSYNTHESE. Show exactly four numbered steps in this order: 1 LICHT, 2 WASSER, 3 CO₂, 4 ZUCKER + SAUERSTOFF. Use a clean editorial science style with one plant cross-section, simple arrows, high contrast, legible labels, and no other text.

GPT Image 2. Provider Together, 1024x1024, n=1, seed=provider random, provider default steps, model ID openai/gpt-image-2, generated Aug 14, 2026, 12:14 PM. Source
Stable Diffusion 3.5 Large. Provider fal.ai, requested 1024x1024, n=1, seed=12166653711354006000, estimated cost $0.068157, model ID fal-ai/stable-diffusion-v35-large, generated Aug 14, 2026, 2:37 PM. Source

Protocol gradually-image-comparison-v1, selection rule: first result.

Prompt SHA 256: ed65cb4e6a51e2336f8b149ad4cbb16e020f07d527ad94362e719d2dcf91d94f

GPT Image 2, image SHA 256: 6c15466e258e2b14a8e862959cb6a67b4f7a23896ea5d42c5a5618d848847711

GPT Image 2, request SHA 256: ed741d3d120188b9286d3576e7dcd3ace9dd7792978d4d5426022f32ed7ca381

Stable Diffusion 3.5 Large, image SHA 256: 4eb0fd741ef42e6d3fee4b1280cdfb098a0a04009b139051963895892e349f1e

Stable Diffusion 3.5 Large, request SHA 256: 8b8014fc63f520d73ff57d3d8df5eae880097383167d44918b963f2830d09940

Product photography

Materials, reflections, and fine details

Prompt and settings

Create a square premium product photograph of a brushed titanium wristwatch standing upright on dark green marble. The watch has a cream dial, thin black hands set to 10:10, twelve distinct hour markers, a realistic crown, and a dark brown leather strap. Soft window light from the left, controlled reflections, shallow depth of field, no text, no logo, no extra objects.

GPT Image 2. Provider Together, 1024x1024, n=1, seed=provider random, provider default steps, model ID openai/gpt-image-2, generated Aug 14, 2026, 12:13 PM. Source
Stable Diffusion 3.5 Large. Provider fal.ai, requested 1024x1024, n=1, seed=13346308312015933000, estimated cost $0.068157, model ID fal-ai/stable-diffusion-v35-large, generated Aug 14, 2026, 2:37 PM. Source

Protocol gradually-image-comparison-v1, selection rule: first result.

Prompt SHA 256: c67fd71e91304458a2feef5c91efa8677260fab0c0099a29a2c52cb579f5b2f3

GPT Image 2, image SHA 256: 471c71b30618afb1a8a67ab8363308d67f99c65ef6740ca72417806d06d3c346

GPT Image 2, request SHA 256: 7b27ff2c9b64529a36bebf2d572c3448f26e68f3e9d1ac6a3f2b5ead8698f60f

Stable Diffusion 3.5 Large, image SHA 256: ea9a7401dee8859ec2cb96af9f7ca29a177ef2202f52f629d8ce9f5839bd4061

Stable Diffusion 3.5 Large, request SHA 256: 21e2c14d1b6315c30134e1390f531349608855d4ef1a87527e6d0ba57b2974ca

Typography and layout

Legible text, hierarchy, and composition

Prompt and settings

Create a square editorial poster for an imaginary night train called MONDFALTER. Show the exact German headline MONDFALTER and the exact subline BERLIN NACH LISSABON. Use a restrained midnight-blue and warm-cream palette, one stylized moth, strong Swiss-grid typography, generous negative space, and no additional words or logos.

GPT Image 2. Provider Together, 1024x1024, n=1, seed=provider random, provider default steps, model ID openai/gpt-image-2, generated Aug 14, 2026, 1:59 PM. Source
Stable Diffusion 3.5 Large. Provider fal.ai, requested 1024x1024, n=1, seed=16632333557910893000, estimated cost $0.068157, model ID fal-ai/stable-diffusion-v35-large, generated Aug 14, 2026, 2:37 PM. Source

Protocol gradually-image-comparison-v1, selection rule: first result.

Prompt SHA 256: 2e012c3d14ee5dd02203a94f394ca80a5038cdebd6032b25a8b8a221f2255419

GPT Image 2, image SHA 256: 434cf9939b9be2a2c2711e13d027c9438c47038bb32229ceeaf405d2e89d2afc

GPT Image 2, request SHA 256: 473a774dfe9f9b7bb597e9e8b63af4e770b6bc4e7f8bc90682636db3516e7353

Stable Diffusion 3.5 Large, image SHA 256: e2994b1883a245b1c8c447957d121ea89d93aafff149029aa4e8507dbb769f9a

Stable Diffusion 3.5 Large, request SHA 256: 530872e813e988ba95025956a3743c6111120af963d90b8970a8bf3eced5547b

How to read this comparison

An Elo score is not a verdict on creativity

A higher arena score means this specific configuration was preferred more often in blind comparisons.

Typography, character consistency, local execution, licensing, and price may matter more to your workload than overall visual appeal.

Prices cover 1,000 images at 1,024 × 1,024 pixels using settings documented by Artificial Analysis. Subscriptions and special rates are excluded.

More matchups

Compare GPT Image 2 and Stable Diffusion 3.5 Large with other important image models under the same evidence rules. Each page shows only exact shared measurements and samples.

View the complete image model comparison

GPT Image 2 vs. Stable Diffusion 3.5 Large

Change image model selection

General specifications and task-specific pricing

Shared capability benchmarks

Shared arena benchmarks

Sample images

Character and detail

Infographic

Product photography

Typography and layout

An Elo score is not a verdict on creativity

Related image model comparisons