GPT Image 2.5 Sunburst vs Nano Banana Pro for Image Tasks
Nano Banana Pro has the edge for typography, attribute binding, advertising composition, garment integration, and coherent relighting, while the remaining scenarios are ties. Choose Nano Banana Pro when those strengths matter most; choose GPT Image 2.5 Sunburst when lower credit cost matters and a tied result is sufficient for your task.
並列資訊比較
| 功能 | GPT Image 2.5 Sunburst | Nano Banana Pro |
|---|---|---|
| 開發商 | OpenAI | |
| Banana Pie 上的供應狀態 | 現已推出 | 現已推出 |
| 所需點數 | 30 點數 | 80 點數 |
| 最高解析度 | 4K | 4K |
| 參考圖像 | 最多 8 張 | 最多 8 張 |
逐情境比較
相同提示詞，兩個模型。從下方發布的圖像親自比較差異。
Product material and lighting
視任務而定The images split the strongest rubric details: the right has the subtler, cleaner reflection, while the left has more consistently modeled upper-left lighting and grounding. Both meet the bottle-count and fill requirements.
- Exactly one transparent rectangular perfume bottle: tie. Each image shows one upright rectangular glass bottle with no visible rendered text, logo, plants, or separate objects; the closures and spray mechanisms are parts of the bottles.
- Half filled with amber liquid: tie. Both bottles have a clearly visible amber liquid level near the midpoint of the glass body, with transparent empty space above.
- Glass refraction and wet-stone reflection: right does better. The right image has one restrained, continuous reflection beneath the bottle, while the left image's reflection spreads into several intense amber and white streaks and is less subtle. Both show visible refraction through the thick glass.
- Upper-left lighting and contact shadow: left does better. Its bright upper-left-facing edges, darker right side, and grounded shadow extending to the right are especially consistent; the right also has upper-left illumination and a contact shadow, but the broad diagonal background shadow competes with the bottle lighting.
使用的提示詞與設定
提示詞
A premium product photograph of exactly one transparent rectangular perfume bottle, half filled with amber liquid, standing upright on wet black stone. Light comes from the upper left, creating coherent refraction, a contact shadow, and one subtle reflection. No text, logo, plants, or extra objects.
Multilingual typography
此項較佳：Nano Banana ProRIGHT satisfies the poster specification more accurately, chiefly because it preserves "MOONLIGHT MARKET" as one line and therefore has exactly four lines.
- Exactly four centered text lines and no additional text: RIGHT does better. It shows four centered lines, while LEFT wraps "MOONLIGHT MARKET" into separate "MOONLIGHT" and "MARKET" lines, producing five lines total.
- Exact text and order: RIGHT does better. Its lines read exactly "MOONLIGHT MARKET", "月光市集", "18 OCT", and "RIVER HALL" in the requested order. LEFT renders "MOONLIGHT" and "MARKET" as two separate lines before "月光市集", "18 OCT", and "RIVER HALL".
- Text colors: Tie. Both render "月光市集" in red and all English and numeric text in black.
- Legibility, spacing, and alignment: RIGHT does better. All four lines are clearly legible, centered, and arranged as a compact, consistently aligned group. LEFT is also legible and centered, but the unintended title wrap disrupts the requested four-line spacing and structure.
使用的提示詞與設定
提示詞
Design a clean cream-colored vertical event poster. Show exactly four centered text lines and no other text: "MOONLIGHT MARKET", "月光市集", "18 OCT", and "RIVER HALL". Render "月光市集" in red and all other lines in black.
Counting and attribute binding
此項較佳：Nano Banana ProBoth outputs follow the requested counts and arrangement closely. The right image has the clearer material distinctions, especially the visible wooden grain and ceramic glaze, giving it the overall edge.
- Exactly three red cubes, two blue spheres, and one yellow mug: Tie. Both images visibly show three cubes in a row on the left, two spheres on the right, and a single mug centered behind them.
- Colors, shapes, and materials: Right does better. Both bind red to cubes, blue to spheres, and yellow to the mug, but the right image shows clearer wood grain on the cubes and a more visibly ceramic, glazed surface on the mug; its spheres also read as transparent glass.
- Mug handle direction: Tie. In both images, the mug handle is clearly attached on and extends toward the right.
- No additional objects or text: Tie. Neither image contains any visible extra objects or rendered text.
使用的提示詞與設定
提示詞
On a matte gray table, exactly three red wooden cubes form a row on the left, exactly two blue glass spheres sit on the right, and one yellow ceramic mug stands centered behind them. The mug handle points right. No other objects or text.
Advertising composition
此項較佳：Nano Banana ProRight follows the text restriction and top-spacing direction more closely while keeping every required element clear and fully visible. Left is visually strong but adds the forbidden text "AERO" and leaves more empty space than requested above the headline.
- Exactly one silver shoe and orange trail: Both show exactly one silver shoe in the lower-right and an orange trail entering from the bottom-left. The right places both elements clearly, while the left shoe is larger and more decisively anchored in the lower-right; slight edge to left.
- Only visible text: Right does better. It shows only "RUN LIGHT" and "42 KM", both spelled correctly. Left also renders "AERO" on the shoe, which violates the requirement for no other visible text.
- Top spacing and headline position: Right does better. Its upper area is visibly empty, with "RUN LIGHT" placed immediately below it in the upper-left. Left also leaves the top uncluttered, but the empty region is substantially deeper before the headline begins.
- Visibility and visual hierarchy: Both keep the required headline, badge, trail, and shoe fully visible. Right has the clearer prompt-compliant hierarchy, with "RUN LIGHT" leading into the prominent "42 KM" badge, trail, and shoe without extra text competing for attention.
使用的提示詞與設定
提示詞
Create a vertical social ad for a fictional running shoe named AERO. Keep the top 15 percent empty. Directly below it, place the headline "RUN LIGHT" in the upper-left. Show exactly one silver shoe in the lower-right, an orange trail curving from the bottom-left, and a round badge reading "42 KM". No other shoes or text.
Structured UI graphic
視任務而定Both images satisfy the title, column, row-label, and button requirements, while both add plan-detail text forbidden by the prompt. Neither has a clear rubric-level advantage.
- Criterion 1: Tie. Both display the exact title "CHOOSE YOUR PLAN" and exactly three columns labeled "STARTER", "PRO", and "TEAM".
- Criterion 2: Tie. Each column in both images contains exactly three aligned labeled rows: "PROJECTS", "STORAGE", and "SUPPORT".
- Criterion 3: Tie. Each column in both images contains exactly one blue button labeled "SELECT".
- Criterion 4: Tie. Both use three evenly aligned columns, but both violate the no-additional-text requirement. LEFT adds "3", "5 GB", "Email", "10", "50 GB", "Priority", "Unlimited", "1 TB", and "24/7". RIGHT adds "5 active projects", "10 GB", "Email support", "20 active projects", "50 GB", "Priority email & chat support", "Unlimited projects", "200 GB", and "24/7 phone, email & chat support".
使用的提示詞與設定
提示詞
Create a clean horizontal pricing comparison graphic titled "CHOOSE YOUR PLAN". Use exactly three equal columns labeled "STARTER", "PRO", and "TEAM". Under each column, show exactly three aligned rows labeled "PROJECTS", "STORAGE", and "SUPPORT", followed by one blue button labeled "SELECT". White background, dark navy text, no additional columns or text.
Six-reference object binding and edit
視任務而定Both outputs visibly satisfy the requested object count and substitutions. Without the six source references, there is no reliable basis for deciding which preserves the exact assets or base scene more faithfully.
- Red mug and white towel replacement: tie. Both images visibly show a dark navy-blue mug and a folded green-and-white striped napkin in comparable tabletop positions, with no red mug or white towel remaining. Exact reference appearance cannot be assessed because References 2 and 3 are not shown.
- Pear and notebook replacement: tie. Both images contain exactly one orange and one mustard-yellow hardcover notebook, and neither visibly retains a green pear or blue notebook. Exact matching to References 4 and 5 cannot be assessed from the outputs alone.
- Two lemons: tie. Both images show exactly two lemons positioned to the right of the mug, with no additional visible fruit or duplicated objects beyond the requested single orange.
- Base-scene preservation: tie. Both retain a wooden tabletop, similar overhead-oblique framing, natural lighting, wood grain, and coherent shadows, but exact preservation of Reference 1's camera, framing, and spatial relationships cannot be verified because the base reference is not shown.
使用的提示詞與設定
提示詞
Use Reference 1 as the base tabletop scene. Replace the red mug with the exact dark navy-blue mug from Reference 2 in the same position and orientation. Replace the folded white towel with the exact green-and-white striped napkin from Reference 3 in the same folded area. Replace the green pear with exactly one orange from Reference 4 in the same position. Replace the blue notebook with the exact mustard-yellow hardcover notebook from Reference 5 in the same position. Add exactly the two lemons from Reference 6 to the right of the mug. Preserve the wooden table, camera, framing, wood grain, lighting, shadows, and all other spatial relationships from Reference 1. Do not copy the white product backgrounds from References 2–6, duplicate any asset, or add other objects.
In-image text replacement
視任務而定Both outputs satisfy the visible text requirements equally well. The preservation criteria require the original image for a reliable comparison, so there is no defensible overall winner.
- Exact text and removal: Tie. Both signs visibly read exactly "NIGHT OWL", with no visible remnants of earlier text.
- Additional or malformed characters: Tie. Neither sign contains extra, duplicated, or malformed characters.
- Font, spacing, and perspective preservation: Cannot be fully assessed because the original image is not shown. Both outputs use widely spaced uppercase sans-serif lettering aligned plausibly with the sign's perspective.
- Material, lighting, and scene preservation: Cannot be fully assessed without the original image. In both outputs, the lettering follows the green painted-wood texture and the storefront lighting appears visually coherent.
參考圖像
使用的提示詞與設定
提示詞
Replace only the sign text with exactly "NIGHT OWL". Preserve the original font style, spacing, perspective, sign material, lighting, and everything else.
Multi-reference identity and garment transfer
此項較佳：Nano Banana ProThe references needed to judge identity and exact garment transfer are absent, so only replacement and physical integration can be compared reliably. On those visible factors, the right output has a small edge for more natural garment fit and drape.
- Identity preservation: Cannot be assessed against Reference 1 because the reference image is not shown here. The two outputs visibly differ in facial proportions, expression, crop, and apparent body width, but there is no visible basis for determining which is closer to the reference.
- Jacket fidelity: Cannot be fully assessed against Reference 2 because the isolated jacket image is not shown here. Both depict light-blue denim, an off-white fleece collar, bronze-colored buttons, flap chest pockets, and a red sleeve patch; the right shows a more naturally open front, while the left shows a fully buttoned construction with somewhat heavier, more regular detailing.
- Full replacement: Tie. Neither output retains a visible black suit jacket, ghost mannequin, or white product background; both place the denim jacket directly on the person in the same neutral interior scene.
- Physical coherence: Right does slightly better. Its open jacket follows the torso and overlaps the white shirt naturally, with plausible folds and sleeve shadows. Both sets of hands are credible, but the left jacket looks unusually rigid and symmetrical around the chest and waist compared with the softer drape on the right.
使用的提示詞與設定
提示詞
Use the portrait in Reference 1 for the person and the isolated jacket in Reference 2 for the garment. Dress the person from Reference 1 in the exact jacket shown in Reference 2. Preserve the person's identity, face, expression, skin, hair, hands, pose, body proportions, background, framing, and lighting from Reference 1. Preserve the jacket's material, color, collar, buttons, pockets, and sleeve patch from Reference 2. Do not copy the ghost mannequin or white product background.
Coherent scene relighting
此項較佳：Nano Banana ProRight provides the more convincing coherent relighting, with clearer localized sunlight and stronger, consistently directed shadows from the left window.
- Warm golden-hour light clearly enters from the left window: Right does better; bright sunlight from the left forms distinct warm window-shaped projections across the rear wall, sofa, rug, and floor. Left also shows a low golden sun at the window, but the room is more uniformly warm.
- Highlights and shadow directions respond coherently to the new light source: Right does better; the plant, table, sofa, and lamp cast pronounced shadows consistently toward the right, away from the left window. Left has directionally plausible table and foliage shadows, but their direction is less consistently legible.
- The result is more than a uniform yellow color filter: Right does better; localized bright projections and deep directional shadows create clear spatial relighting. Left has localized highlights and shadows too, but the pervasive orange cast makes it look somewhat closer to a broad warm filter.
- No furniture or decor is moved, added, removed, or redesigned: This cannot be assessed reliably without the original source image. The two outputs visibly differ in framing and in details of the window, sofa, table, rug, plant, and lamp, but that does not establish which one preserved the source more faithfully.
參考圖像
使用的提示詞與設定
提示詞
Change the lighting to warm golden-hour sunlight entering from the left window. Do not move, add, remove, or redesign any object. Update highlights and shadows coherently.
Cross-ratio outpainting
視任務而定Both outputs appear visually plausible as landscape extensions, and the only clearly assessable extension-quality criterion is effectively tied. Preservation of the original cannot be judged without seeing the source image.
- 16:9 output and natural side content: Tie. Both outputs are wide landscape compositions with plausible beach, ocean, and sky continuing to both sides; exact aspect ratio cannot be reliably confirmed from the displayed images alone.
- Complete original centered without cropping or stretching: Tie. The cabin is centered and shows no obvious stretching in either output, but the original source image is not shown, so completeness and cropping cannot be verified.
- Original cabin, shoreline, and internal composition unchanged: Cannot be assessed for either side because the original image is absent. The cabin and shoreline differ between the two outputs, but there is no visible reference establishing which, if either, preserved the source.
- No visible seams, mirrored filler, or repeated objects: Tie. Neither output shows an obvious boundary seam, mirrored region, or conspicuously repeated object; the extended sand, waves, and clouds blend naturally on both sides.
參考圖像
使用的提示詞與設定
提示詞
Expand the canvas to a 16:9 landscape by naturally continuing the beach, ocean, and sky on both sides. Keep the complete original image centered without cropping, stretching, letterboxing, or modifying it.
Human anatomy and contact
使用的提示詞與設定
提示詞
A photorealistic close-up of an adult potter shaping a clay bowl on a spinning wheel. Both hands are fully visible, each with five natural fingers touching the clay. No other people or hands. Soft window light.
Product preservation across scene change
參考圖像
使用的提示詞與設定
提示詞
Place the sneaker on a wet outdoor basketball court at dusk. Preserve the exact sneaker shape, black geometric side mark, materials, stitching, sole geometry, and camera angle. Add physically coherent contact, reflections, and dusk lighting.
成本與生成時間
GPT Image 2.5 Sunburst starts at 30 credits, while Nano Banana Pro starts at 80, making Sunburst the lower-credit option. In this small run, observed latency ranged from about 69 s to 155 s per scenario for Sunburst and from about 39 s to 68 s for Nano Banana Pro; these are run-specific samples, not benchmark figures.
比較方式與揭露事項
We used one fixed prompt per scenario and made one run with each model per scenario, publishing the outputs as generated. Banana Pie sells paid access to both models.
Banana Pie 在同一個創作台提供此模型及其他模型的付費使用權。我們的評測結論來自與使用者相同生成流程的測試。