OpenAI
Generate ad creatives on Martini using GPT Image 2 — strong reasoning over compositional briefs and multi-element prompts, with reasonable in-image text rendering for short headlines and CTAs. GPT Image 2 sits between Ideogram (typography champ) and FLUX.2 (photographic prompt fidelity): it understands narrative briefs better than either, which makes it the right pick for ad concepts that require interpretation ("show a busy parent finding 5 minutes for themselves") rather than literal staging. Pair it with a downstream Ideogram node for the final headline pass when typography needs to be perfect.
GPT Image 2 reasons over scenarios. Write: "A busy parent in their late thirties, momentarily alone in a sunlit kitchen, taking a deep breath while holding a coffee mug — convey a sense of hard-earned calm." GPT Image 2 will compose the scene including subtle storytelling details. Compare to FLUX.2, which would render exactly what you describe but not infer the emotional context.
GPT Image 2 renders short in-image text reliably — single headlines, CTA buttons, brand marks. Quote the text in the prompt: "with the headline '5 MINUTES IS ENOUGH' in clean sans-serif, upper third." For long-form copy, multi-line headlines, or fine typography, chain to Ideogram. GPT Image 2 handles the basics; Ideogram is the production tool when text fidelity is critical.
GPT Image 2 handles 5+ named elements in one prompt without losing track ("a runner, a coffee cup on a bench, a sunrise over a coastal trail, athletic shoes prominently visible, the headline 'TRAIN HARDER'"). Other Martini image models start dropping elements past 3-4 named items. Use GPT Image 2 when the ad brief is dense with required visual elements.
Duplicate the GPT Image 2 node per aspect — 1:1, 4:5, 9:16, 16:9 — and adjust the compositional framing in the prompt per node. Unlike FLUX.2, GPT Image 2 may compose less consistently across pure aspect changes; explicitly direct the framing per ratio ("vertical composition with subject offset right for 9:16," "centered tight crop for 1:1").
For ads where the headline must be pixel-perfect (paid social legibility, App Store screenshots, OOH), generate the photographic layer with GPT Image 2 (no in-image text), then route into a downstream Ideogram V3 node that adds the headline + CTA. GPT Image 2 + Ideogram = narrative composition + production typography. Two nodes, one canvas.
Once a concept tests well, save the canvas — GPT Image 2 photo node, downstream Ideogram nodes, aspect-ratio fan-out, parameter pins. Next campaign, swap the narrative brief and the headline string. The ad pipeline becomes reusable: same template structure, different campaign, repeatable production cadence.
Narrative ad brief with built-in text. GPT Image 2 reasons through the emotional context and renders short text reliably.
A busy parent in their late thirties, momentarily alone in a sunlit kitchen, taking a deep breath while holding a coffee mug. Soft morning light from a window on camera left. Convey a sense of hard-earned calm. The headline "5 MINUTES IS ENOUGH" in clean sans-serif on the upper third. CTA "TRY 14 DAYS FREE" lower right. 1:1 aspect.
GPT Image 2 generates the photographic layer; Ideogram adds the typography downstream. Cleaner production pipeline for high-stakes campaigns.
Same parent concept, vertical composition for TikTok with subject offset to the right third for headline space on the left, soft morning palette, no text in this output (text added downstream by Ideogram), 9:16 aspect.
GPT Image 2 handles 6+ named elements without dropping any. The dense composition is where it outperforms FLUX.2 and Midjourney.
Dense multi-element ad brief: a runner mid-stride, a coffee cup steaming on a wooden bench in the foreground, sunrise over a coastal trail in the background, athletic shoes prominently visible in the lower third, the headline "TRAIN HARDER" in bold sans-serif top-third, CTA button "SHOP NOW" lower right, 1:1 aspect.
Narrative + emotional context + serif typography. GPT Image 2 reasons through the story; quoted serif holds for production.
Aspirational lifestyle: a graduate in cap and gown standing on a campus quad, late afternoon golden light, holding a folded diploma, classmates blurred in the background, conveying achievement and possibility. The headline "READY FOR WHATEVER NEXT" in confident serif, lower third. 4:5 aspect.
Use GPT Image 2 when the brief is narrative ("show a parent finding peace") rather than literal ("a runner on a trail"). For literal staging, FLUX.2 is the cleaner pick.
Quote in-image text. GPT Image 2 renders quoted strings reliably for short copy; long-form headlines should chain to Ideogram.
Lean on multi-element composition — GPT Image 2 handles 5-7 named elements per prompt without dropping items. This is its differentiator vs FLUX.2 / Midjourney.
For variant testing, change the narrative emphasis ("hard-earned calm" → "joyful release") rather than swapping single keywords. GPT Image 2 reads holistically.
Set quality to "high" for hero ads, "standard" for variants and exploration. Quality affects both detail and prompt adherence.
For pixel-perfect typography in production ads, generate the photo with GPT Image 2 (no in-image text) and chain to Ideogram for the final overlay.
GPT Image 2 returns 1024-2048 wide outputs with strong narrative coherence and reasonable in-image text rendering for short copy. Generation time 30-60s per output. Best at multi-element compositions and brief interpretation; pair with Ideogram for typography-critical work and FLUX.2 for literal product staging. Output drops onto the canvas for downstream chaining — sequence builder for animated ads, NLE export for native delivery to ad platforms.
Connect GPT Image 2 with other AI models on Martini's infinite canvas. No GPU required — start free.
Get Started FreeIdeogram
Fan one ad concept into a 30-asset paid-social matrix on Martini using Ideogram V3 — every aspect ratio, every CTA variant, every headline localization, with the copy rendered legibly inside the image. Ideogram is the only AI image model that handles in-image text reliably; for performance-marketing work where the headline IS the asset, no other model in the stack comes close. Pair it with a saved brand prompt, fan into ratio-specific nodes, and ship a Meta + TikTok + YouTube test matrix without a Figma round-trip.
View guideBlack Forest Labs
Generate the visual layer of an ad-creative matrix on Martini using FLUX.2 — lifestyle hero photography, product context shots, and aspirational scenes that feed downstream Ideogram nodes for headline overlay. FLUX.2 is the prompt-fidelity workhorse: it renders specific products, environments, and compositions almost literally, which is exactly what performance marketers need when the creative brief reads like a shot list. For text overlays, chain to Ideogram; for the photography itself, FLUX.2 is the cleaner pick.
View guide