Neon Honey Rush
Material study · Generated with GPT Image 2
Describe the scene in natural language. GPT Image 2 turns it into high-resolution artwork, then places the result on a Grida canvas where you can keep working.
- Input
- Natural language
- Default
- 1024×1024
- Speed
- ~1m
Outputs
One model. Very different directions.
These are real GPT Image 2 outputs made for Grida’s Lyria music showcase. The same model moved from product-like material studies to cinematic environments, interiors, and abstract concept art.
Horizon Of Light
Fantasy environment
Lamplight Letters
Editorial interior
Fracture Protocol
Concept art
Petals In Static
Dreamlike photography
Midnight Ink Cipher
Cinematic still life
Golden Hour Groove
Editorial fashion
Each cover and its companion Lyria track began from one shared creative brief. No stock photography is used in this gallery. The stored web outputs are 1024×1024; original quality, seed, and provider request identifiers were not retained.
Direction
One direction, two kinds of output.
This retained brief was written to shape a companion track. Its pace, instrumentation, and late-night mood also became the direction for the cover, showing how GPT Image 2 can carry an art direction across media.
“Confident, driving retro-electronic synthwave at 104 BPM in F# minor; gated reverb drums, analog bass, lush JP-8 pads, chorused FM lead, and a lead male vocal with subtle tape saturation and 1/8-dotted delay. Slightly melancholic, cinematic, late-night highway energy.”
Excerpt from the shared brief behind “Neon Mile Marker.” The same source prompt directed both its Lyria track and GPT Image 2 cover artwork. For a direct image request, be more explicit about the subject, composition, materials, lighting, and any words that must appear.
Read the complete creative brief
Confident, driving retro-electronic synthwave at 104 BPM in F# minor; gated reverb drums, analog bass, lush JP-8 pads, chorused FM lead, and a lead male vocal with subtle tape saturation and 1/8-dotted delay. Slightly melancholic, cinematic, late-night highway energy. [Intro] [00:00 - 00:18] Arpeggiated saw synth, slow-build pad swell, distant tom fills, side-chained kick enters at bar 5. No vocals. [Verse 1] [00:18 - 00:45] Dry verb on vocal, muted palm-bass, soft hi-hats. "Headlights cut the fog like a confession, radio static humming through my chest. I traced your name in the dashboard's reflection, eighty miles from anywhere I'd rest." [Chorus] [00:45 - 01:10] Big gated snare, stacked vocal harmonies a 5th up, bright lead synth countermelody. "Run me through the neon, mile after mile, I'm a ghost in a jacket, learning how to smile. Somewhere past the static, somewhere past the rain, I'll forget your number, but I won't forget the name." [Verse 2] [01:10 - 01:35] Add shimmering bell synth, light arpeggio panning. "Every exit sign is a promise I outran, every diner window holds a younger man. The moon's a cassette I keep rewinding slow, playing back the places I was scared to go." [Chorus] [01:35 - 02:00] Full arrangement, doubled lead vocal, octave bass. Repeat chorus lyrics with ad-lib "oh-oh" tag. [Bridge] [02:00 - 02:18] Drums drop to filtered loop, vocoder pad, half-time feel. "Maybe love's just a frequency, fading in, fading out of me." [Outro] [02:18 - 02:40] Full band returns, lead synth solo over chorus chords, long tape-delay tail, slow fade.
Composition
Three frames, ready in the playground.
Grida currently exposes square, portrait, and landscape presets. The upstream model supports a wider custom-size envelope, but we keep that capability separate from controls that are actually playable today.
- Preset outputs
- 1024×1024 · 1024×1536 · 1536×1024
- Upstream size envelope
- 0.655–8.29 MP
- Published output price
- $0.005–$0.211 / image
- Catalogue ID
- openai/gpt-image-2
Published pricing
Output price follows quality and frame.
These are OpenAI’s published image-output prices for the three presets. OpenAI may also charge for prompt or reference-image input tokens. Grida’s current playground exposes size—not a quality picker—and prices the request against the medium output tier for the selected preset.
Swipe or scroll horizontally to compare every size.
| Quality | 1024×1024 | 1024×1536 | 1536×1024 |
|---|---|---|---|
| Low | $0.006 | $0.005 | $0.005 |
| Medium | $0.053 | $0.041 | $0.041 |
| High | $0.211 | $0.165 | $0.165 |
In Grida
Generate, then continue on the canvas.
The playground keeps GPT Image 2 selected, generates through Grida-hosted AI, and places the result directly into a canvas. Authentication and organization credit are requested only when execution requires them.
Try GPT Image 2Describe
Write the subject, mood, materials, composition, and light.
Generate
Choose a frame and run the exact GPT Image 2 model.
Continue
Inspect the output on a real Grida canvas and keep designing.
FAQ
GPT Image 2 in Grida
The practical details before you generate.
What is GPT Image 2?
GPT Image 2 is OpenAI’s image generation and editing model. It accepts text and image input, produces image output, and supports flexible image sizes. Grida’s current playground exposes its text-to-image generation path.
What image sizes can I generate in Grida?
The playground offers 1024×1024, 1024×1536, 1536×1024 presets. GPT Image 2 also supports custom dimensions in 16-pixel steps inside its published 0.655–8.29 megapixel envelope, up to a 3:1 aspect ratio.
How much does a GPT Image 2 generation cost?
Published image-output pricing currently ranges from $0.005–$0.211 per image depending on quality and size. The current Grida playground debits the selected preset at its catalogue medium-tier amount from your organization’s available AI credit; access is based on entitlement, not on how that credit was granted.
Can I edit an existing image with GPT Image 2 in Grida?
The upstream model supports image input and editing. The public Grida demo on this page currently supports text-to-image generation only, so we do not present editing controls that are not yet wired.
Give the next image a clear direction.
Open the playground with GPT Image 2 already selected. Sign in and organization credit are handled at generation time.
Open GPT Image 2 playground