Head-to-head comparison

GPT Image 2 vs. GPT Image 1.5

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: 48
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 GPT Image 1.5
Attribute	GPT Image 2	GPT Image 1.5
Provider	OpenAI	OpenAI
Released	April 2026	December 2025
Access	API, Web app	API, Web app
Lifecycle	No separate lifecycle notice	Marked as deprecated by OpenAI; removed from the API on December 1, 2026 Status source
Weights	Proprietary	Proprietary
Parameters	Not published	Not published
Maximum output	4K	1,536 × 1,024 pixels
Arena configuration	High	High
Text-to-image price	$211 per 1,000 images	$133 per 1,000 images
Editing price	$211 per 1,000 images	$133 per 1,000 images
Technical source	Model sourceVerified 2026-08-15	Model sourceVerified 2026-08-15

GPT Image 2GPT Image 1.5

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

GPT Image 2$211

GPT Image 1.5$133

Image Editing ArenaUSD per 1,000 images at 1,024 × 1,024 pixels

GPT Image 2$211

GPT Image 1.5$133

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. This pair also shares 32 results from published research benchmarks.

GPT Image 2GPT Image 1.5

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

GPT Image 241.4 / 100

GPT Image 1.529.9 / 100

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

GPT Image 283.9 / 100

GPT Image 1.543.7 / 100

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

GPT Image 293.1 / 100

GPT Image 1.577 / 100

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

GPT Image 278.2 / 100

GPT Image 1.542.5 / 100

Key to the marks:better valuebetter, but not conclusivetie

Capability benchmark scores for GPT Image 2 and GPT Image 1.5
Benchmark	GPT Image 2	GPT Image 1.5
GenExam Mathematics, strict	Higher benchmark score50.3 %	26.5 %
GenExam Physics, strict	Higher benchmark score79.6 %	46 %
GenExam Chemistry, strict	Higher benchmark score69.5 %	39 %
GenExam Biology, strict	Higher benchmark score89.1 %	56.4 %
GenExam Geography, strict	Higher benchmark score84.8 %	60.6 %
GenExam Computer science, strict	Higher benchmark score73.5 %	36.3 %
GenExam Engineering, strict	Higher benchmark score79.3 %	44.1 %
GenExam Economics, strict	Higher benchmark score83.1 %	42.9 %
GenExam Music, strict	Higher benchmark score64.6 %	29.2 %
GenExam History, strict	Higher benchmark score82.9 %	51.2 %
GenExam overall, strict	Higher benchmark score74.6 %	43.2 %
GenExam Mathematics, relaxed	Higher benchmark score85.2 %	65.8 %
GenExam Physics, relaxed	Higher benchmark score95.6 %	85.4 %
GenExam Chemistry, relaxed	Higher benchmark score92 %	78.1 %
GenExam Biology, relaxed	Higher benchmark score97.5 %	91.9 %
GenExam Geography, relaxed	Higher benchmark score97.6 %	92.5 %
GenExam Computer science, relaxed	Higher benchmark score93.3 %	75.8 %
GenExam Engineering, relaxed	Higher benchmark score96.5 %	86.4 %
GenExam Economics, relaxed	Higher benchmark score97.7 %	85.5 %
GenExam Music, relaxed	Higher benchmark score89.1 %	70.8 %
GenExam History, relaxed	Higher benchmark score97.1 %	90.9 %
GenExam overall, relaxed	Higher benchmark score93.8 %	82.3 %
Qwen Image Bench quality	Higher benchmark score58.65 / 100	55.14 / 100
Qwen Image Bench aesthetics	Higher benchmark score67.53 / 100	60.88 / 100
Qwen Image Bench alignment	Higher benchmark score65.85 / 100	61.72 / 100
Qwen Image Bench real-world fidelity	Higher benchmark score57.38 / 100	53.95 / 100
Qwen Image Bench creative generation	Higher benchmark score75.23 / 100	66.35 / 100
Qwen Image Bench overall	Higher benchmark score64.69 / 100	59.65 / 100
GRADE reasoning	Higher benchmark score82.2 / 100	54.5 / 100
GRADE consistency	Higher benchmark score94.4 / 100	82.3 / 100
GRADE readability	Higher benchmark score98.8 / 100	90.7 / 100
GRADE accuracy	Higher benchmark score56 / 100	16 / 100
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	29.9 / 100 Mean rank: 21.33 (23 / 22 / 19) Five rubric scores Headline: 1.67 / 2 Subline: 1.67 / 2 Text control: 1.67 / 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	43.7 / 100 Mean rank: 17.33 (22 / 8 / 22) Five rubric scores Subject: 1.67 / 2 Time: 1 / 2 Watch details: 1.67 / 2 Photography: 1.67 / 2 Exclusions: 1.67 / 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	77 / 100 Mean rank: 7.67 (6 / 5 / 12) 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	42.5 / 100 Mean rank: 17.67 (20 / 14 / 19) Five rubric scores Title: 1.67 / 2 Sequence: 1.67 / 2 Labels: 1 / 2 Diagram: 1.67 / 2 Clarity: 1 / 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.

12 shared result series

GPT Image 2GPT Image 1.5

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

GPT Image 21,171 Elo

GPT Image 1.51,102 Elo

Image Editing ArenaBlind preference from the same prompt, higher is better

GPT Image 21,122 Elo

GPT Image 1.51,103 Elo

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

GPT Image 21,361.089

GPT Image 1.51,214.044

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

GPT Image 21,365.89

GPT Image 1.51,224.786

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

GPT Image 21,399.599

GPT Image 1.51,245.019

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

GPT Image 21,389.09

GPT Image 1.51,240.235

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

GPT Image 21,381.149

GPT Image 1.51,239.16

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

GPT Image 21,380.609

GPT Image 1.51,249.33

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

GPT Image 21,427.407

GPT Image 1.51,261.708

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

GPT Image 21,426.635

GPT Image 1.51,253.437

LMArena Image Editing: Multi-image editingBlind preference from the same prompt, higher is better

GPT Image 21,453.901

GPT Image 1.51,342.523

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

GPT Image 21,461.047

GPT Image 1.51,370.19

Key to the marks:better valuebetter, but not conclusivetie

Arena benchmark scores for GPT Image 2 and GPT Image 1.5
Benchmark	GPT Image 2	GPT Image 1.5	Source
Text to Image Arena	Higher score1,171 Elo Rank 3, 95% confidence interval: ±9, 15,176 ratings	1,102 Elo Rank 9, 95% confidence interval: ±9, 14,748 ratings	Artificial AnalysisRetrieved 2026-09-22
Image Editing Arena	Higher score1,122 Elo Rank 5, 95% confidence interval: ±9, 13,741 ratings	1,103 Elo Rank 12, 95% confidence interval: ±9, 12,789 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	1,214.044 Rank 15, 95% confidence interval: ±6.6, 15,526 ratings	LMArenaRetrieved 2026-09-19
LMArena Text to Image: Art	Higher score1,365.89 Rank 2, 95% confidence interval: ±7.91, 10,214 ratings	1,224.786 Rank 14, 95% confidence interval: ±5.9, 21,424 ratings	LMArenaRetrieved 2026-09-19
LMArena Text to Image: Cartoon	Higher score1,399.599 Rank 3, 95% confidence interval: ±6.48, 31,995 ratings	1,245.019 Rank 14, 95% confidence interval: ±4.72, 59,294 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	1,240.235 Rank 14, 95% confidence interval: ±4.6, 58,887 ratings	LMArenaRetrieved 2026-09-19
LMArena Text to Image: Overall	Higher score1,381.149 Rank 3, 95% confidence interval: ±4.38, 78,731 ratings	1,239.16 Rank 15, 95% confidence interval: ±3.24, 151,702 ratings	LMArenaRetrieved 2026-09-19
LMArena Text to Image: Photorealism	Higher score1,380.609 Rank 3, 95% confidence interval: ±6.29, 30,097 ratings	1,249.33 Rank 15, 95% confidence interval: ±4.7, 60,842 ratings	LMArenaRetrieved 2026-09-19
LMArena Text to Image: Portraits	Higher score1,427.407 Rank 3, 95% confidence interval: ±8.27, 15,130 ratings	1,261.708 Rank 12, 95% confidence interval: ±5.81, 29,731 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	1,253.437 Rank 16, 95% confidence interval: ±4.84, 54,527 ratings	LMArenaRetrieved 2026-09-19
LMArena Image Editing: Multi-image editing	Higher score1,453.901 Rank 3, 95% confidence interval: ±4, 101,345 ratings	1,342.523 Rank 11, 95% confidence interval: ±3.62, 192,446 ratings	LMArenaRetrieved 2026-09-19
LMArena Image Editing: Overall	Higher score1,461.047 Rank 3, 95% confidence interval: ±3.45, 235,928 ratings	1,370.19 Rank 15, 95% confidence interval: ±2.81, 566,510 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
GPT Image 1.5. Provider Together, 1024x1024, n=1, seed=provider random, provider default steps, model ID openai/gpt-image-1.5, generated Aug 14, 2026, 12:17 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

GPT Image 1.5, image SHA 256: 0dff52ccd39400ac66ac8870c9ea937512ff1474688e70118688f420c38c2296

GPT Image 1.5, request SHA 256: 5199631ac6ff07633dd2fb6ab61c918c26e1e02f37e6131828072233ac04527d

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
GPT Image 1.5. Provider Together, 1024x1024, n=1, seed=provider random, provider default steps, model ID openai/gpt-image-1.5, generated Aug 14, 2026, 12:17 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

GPT Image 1.5, image SHA 256: 1c9ff59c32621c8b48b002dacb0ae66487262280c7c0a5b3d6456c81b248ebe9

GPT Image 1.5, request SHA 256: 7f1efc9531f1e6bed7d61a11c637add9fd892001aea2bb2781f8d89c79a2eb03

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.

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

GPT Image 1.5, image SHA 256: 67cfa2f324447027ceb5f82e9446d40a1395b46ce425890a74606b4d97731cf6

GPT Image 1.5, request SHA 256: c850843a298631a4076c5d2686de8e9d7223534f2b8ca53505f437862718bef9

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
GPT Image 1.5. Provider Together, 1024x1024, n=1, seed=provider random, provider default steps, model ID openai/gpt-image-1.5, generated Aug 14, 2026, 1:59 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

GPT Image 1.5, image SHA 256: 6a4cec86850fd371b65274198e2bf630b3cf36a096f9fb953d4bfe17c3222ebe

GPT Image 1.5, request SHA 256: 9c3a396aecfcaae3f8a25edbe5c8c44d0d33199d7ba29efda864d2b0221c3250

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 GPT Image 1.5 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. GPT Image 1.5

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