Qwen Image 3 vs Nano Banana 2.1: Same-Prompt Test

All guides · Best AI image models

Quick answer: across six identical tests in Fuser, Qwen Image 3 won three, Nano Banana 2.1 (now the default model in Fuser's Gemini Image node) won two and one was a tie. Qwen Image 3 was the more literal model: it kept a café menu flat, changed only the background in a product edit and kept every character on the label, down to its small dots. Nano Banana 2.1 set a poster exactly with nothing added and gave a photoreal portrait the soft background blur the prompt asked for, but misspelled two small words on the label during an edit. It is also the cheaper of the two in Fuser: at 2K, Qwen Image 3 costs about one and a half times as much. Both run as nodes in Fuser, so you can wire one prompt to each and keep the version you would ship. Each test is one run per model, so read this as tendencies, not a benchmark.

How we tested

We reran the prompts and the product photo from our earlier same-prompt tests (FLUX.2 vs Nano Banana, Nano Banana vs GPT Image and best AI models for text in images) on both models on 6 October 2026, the day Google released Nano Banana 2.1. This page previously tested Nano Banana 2, which Google shuts down on 29 October 2026 (what changed in Nano Banana 2.1). Every output here is a single run; nothing was re-rolled or picked from several takes.

• Models: Qwen Image 3 (Alibaba's Qwen-Image-3.0, the model on Fuser's Qwen Image 3 node) and Nano Banana 2.1 (Google's gemini-nano-banana-2.1, the default model on Fuser's Gemini Image node). In Fuser the Gemini Image node calls Google directly; for this test we generated both models' images with the same model versions Fuser runs.

• Resolution: 2K for both, the highest setting on Fuser's Qwen Image 3 node and the default on its Gemini Image node. Qwen Image 3 rendered at exactly what its node sends (2048 × 1536 at 4:3, 2048 × 2048 square, 1536 × 2048 portrait). Nano Banana 2.1 returned 2400 × 1792 at 4:3, 2048 × 2048 square and 1792 × 2400 at 3:4.

• Aspect ratios: Qwen Image 3's node offers square, 4:3 and 16:9 shapes, so the photo test used 4:3 for both models. The posters and menus used 3:4 portrait.

• Other settings: each node's defaults. For Qwen Image 3 that means Expand Prompt on (Alibaba lists its prompt rewriting as on by default and "recommended" in the API reference); we fixed its seed at 4242. Fuser does not send a seed to Nano Banana 2.1.

• Judging: we read every output at full resolution, counted the requested characters on the poster and menu (spaces excluded: 111 and 172), counted objects and compared edited labels with the source letter by letter.

Test 1: a photoreal portrait

Prompt: "A photorealistic candid photograph of an elderly fisherman mending a bright orange net on a weathered wooden dock at dawn. Mist hangs over the harbour behind him, his calloused hands are in sharp focus, and the background falls off into soft blur. Natural light, 50mm lens look, muted colours."

• Qwen Image 3 set a cool, misty morning with a lighthouse in the distance, kept the calloused hands sharp around a wooden netting needle and filled the dock with props: a bucket of rope, a wooden chest, a spool of twine and scissors. The fishing boats behind him stay fairly crisp, so the mist does the softening rather than the lens.

• Nano Banana 2.1 produced the larger image (2400 × 1792) in the same cool, misty light, with sharp hands and a clearly out-of-focus harbour behind him. The net is a saturated orange. Neither model lettered anything into the scene this time.

Verdict: Nano Banana 2.1, narrowly. Its background "falls off into soft blur" as written. Neither image shows a sunrise; both read as an overcast, misty morning rather than dawn. Nano Banana 2 painted boat names on this prompt in two earlier tests; Nano Banana 2.1 added none in this run.

Test 2: counted objects in set positions

Prompt: "Overhead flat lay on a grey linen tablecloth: exactly three lemons on the left, a blue enamel mug in the center, two croissants on a white plate on the right, and a folded newspaper along the top edge. Natural morning light, photorealistic."

• Qwen Image 3 placed three lemons on the left, a blue enamel mug of black coffee in the middle, two croissants on a white plate on the right and a newspaper across the top. It invented a correctly spelled masthead and headline, "THE DAILY HERALD" and "Global Markets See Steady Growth", over unreadable body text.

• Nano Banana 2.1 got the same counts and positions, with a milky coffee in the mug and the newspaper folded into the top-right corner. Its newsprint is unreadable pseudo-text.

Verdict: a tie. Both followed every count and placement. Both invented newspaper text that was not in the prompt, which is worth checking if the image will be seen up close.

Test 3: a poster and a menu with exact text

Poster prompt: a vertical jazz-night poster with the headline "NIGHT BLOOMS", the line "Live at the Harbor Room", a flat illustration of a brass trumpet among white moonflowers on deep teal, and three lines of small cream sans-serif text: the date "Friday, October 17", a doors-and-tickets line, and "With the Mara Quintet and special guest Theo Lindqvist". The full prompt is in our text-in-images roundup; we used it verbatim.

Menu prompt: "A printed café menu card in flat graphic design: cream paper, dark brown ink, a small coffee cup icon. Title at the top: "CORNER STORE COFFEE". Below it, a price list with exactly these eight items and prices, one per line: "Espresso 3.20", "Cortado 3.80", "Flat White 4.10", "Oat Latte 4.60", "Cardamom Bun 3.90", "Rye Sourdough Toast 5.40", "Pistachio Croissant 4.75", "Iced Hojicha 4.95". At the bottom, in small text: "Open daily 7:30 to 16:00, 14 Wexford Lane"."

• Poster, Qwen Image 3: 111 of 111. Every character is right, including the hyphen in the "Doors 8 PM - Tickets" line, with the headline split over two lines. It also added a "HARBOR ROOM JAZZ CLUB" logo and small sparkles that nobody asked for.

• Poster, Nano Banana 2.1: 111 of 111, with a tall condensed headline and nothing added.

• Menu, Qwen Image 3: 172 of 172. All eight items, all eight prices and the footer were exact, on a flat card as the prompt asked, with dot leaders and printer's corner marks.

• Menu, Nano Banana 2.1: 172 of 172, each line once, with dot leaders. It showed the card as a print lying on a wooden table rather than flat graphic design.

Verdict: split. Nano Banana 2.1 for the poster, Qwen Image 3 for the menu. Both spelled everything right; the difference is what each added around the text. Alibaba says Qwen-Image-3.0 "supports precise rendering of text as small as 10px" and accepts up to 4.5k tokens of input (Alibaba Cloud); Google lists "enhanced text rendering and infographic layout accuracy" among Nano Banana 2.1's updates (model card). For more models on the same two prompts, see best AI models for text in images.

Test 4: change the background, keep the product

Both models got the same 1024 × 1024 product photo, a frosted hand wash bottle with small type on its label, and the instruction "Change the background to a terrazzo counter."

• Qwen Image 3 replaced the grey backdrop with a terrazzo wall and counter and changed nothing else: same framing, same bottle, no props.

• Nano Banana 2.1 stood the bottle on a terrazzo counter in front of a pale tiled wall and added no props, but it redrew the label's small print.

The label close-ups decide it. Qwen Image 3 kept every character, including the small dots in "GENTLE · NATURAL · EFFECTIVE". Nano Banana 2.1 printed that line as "EONELE - NATURAL - SFFECTIVE": two misspelled words, with the dots turned into hyphens. Nano Banana 2 made a similar two-word slip on this same edit in our FLUX.2 comparison, so small print through an open-ended edit is still the place to proofread.

Verdict: Qwen Image 3. It did only what was asked and left the packaging text exactly as it was.

Test 5: replace one line of label text

Same photo, new instruction: "Change the line "BERGAMOT & SAGE" on the label to "CEDAR & FIG". Keep every other word and number on the label, the bottle, the pump and the background exactly as they are."

Both models swapped the line to "CEDAR & FIG" and kept "SOLA", "HAND WASH", the three label words and "300 ml ℮ 10.1 fl oz", with the bottle, pump and grey background unchanged. Qwen Image 3 kept the middle dots; Nano Banana 2.1 again turned them into hyphens (the last two crops above).

Verdict: Qwen Image 3, narrowly. With an explicit "keep everything else" instruction, Nano Banana 2.1 spelled every word right this time, which matches Google's own edit example in the Gemini API docs: "Do not change any other elements of the image." More text-edit technique in our Qwen Image 3 guide.

What each model offers beyond this test

• Qwen Image 3: Alibaba announced Qwen-Image-3.0 in July 2026 as the third generation of its Qwen-Image series, citing inputs of up to 4.5k tokens, legible text down to 10px and native rendering in 12 languages (Alibaba Cloud). Its API produces between 512 × 512 and 2048 × 2048 total pixels and edits from 1 to 3 input images (API reference). Fuser's node offers 1K (the default) or 2K, six shapes (square, square HD, 4:3 and 16:9 in portrait or landscape), up to 3 reference images (connecting one switches it to editing), a negative prompt, Expand Prompt and a seed.

• Nano Banana 2.1: Google released Nano Banana 2.1 (gemini-nano-banana-2.1) on 6 October 2026 as an update to Nano Banana 2, citing improved visual quality, prompt adherence, multi-turn character consistency and text rendering (model card). It offers 1K, 2K and 4K output, up to 14 reference images (consistency for up to 4 characters and 10 objects), Google Search grounding, configurable thinking and a SynthID watermark on every image (Gemini API docs). Fuser's Gemini Image node offers 1K, 2K (the default) or 4K, Auto or ten aspect ratios, up to 10 input images, a system prompt and a Search Grounding toggle, and the same node also runs Nano Banana Pro, Nano Banana 2 Lite and the original Nano Banana.

Relative cost

Both nodes charge per image, and the price depends on resolution. In Fuser's credits:

• At 2K (as tested): Qwen Image 3 costs about one and a half times what Nano Banana 2.1 does.

• At 1K: Qwen Image 3 costs about 1.2 times Nano Banana 2.1.

• Above 2K: only Nano Banana 2.1 goes further; its 4K output costs about 2.25 times its 2K price.

Which one to use

• Pick Qwen Image 3 for price lists, menus and product edits where everything you did not mention has to stay exactly as it was, down to the punctuation on a label. Check its output for logos or badges it adds on its own.

• Pick Nano Banana 2.1 for the cheaper 2K image, for 4K, for an aspect ratio Fuser's Qwen Image 3 node does not offer (3:2, 4:5, 21:9), for more than three reference images, or for search grounding on real-world subjects. Proofread small print after an edit.

• Run both when the text has to be right. In Fuser, wire one prompt node into a Qwen Image 3 node and a Gemini Image node, as in the canvas at the top, and keep the version that needs fewer fixes. For fine print that must be exact every time, set it as real text in the Compositor instead. See also our poster typography workflow.

Related: Qwen Image 3 vs Qwen Image 2, Nano Banana prompt guide, Seedream vs Nano Banana and best AI image editing models.