MOSS-Image vs Qwen-Image
Both are open models whose makers lead with text rendering: Alibaba's 20 billion parameter Qwen-Image from August 2025, and OpenMOSS's 5 billion parameter MOSS-Image from 16 October 2026, which plans a JSON layout before drawing. Both ran the same twelve prompts through fal at the same sizes: Qwen-Image on 10 September, MOSS-Image on 22 October. Neither has an edit endpoint on fal.
Anvisha Pai, Co-founder & CEO, Voyager
Which to use
MOSS-Image for anything with small text or structure. Its chart is correct, values, months and bar heights, where Qwen-Image's month labels read 'Muy', 'January', 'The', 'Months'. Its infographic spelled every step, where Qwen-Image's title read 'pour-oyf' and step 3 dissolved into marks. Its logo is a cleaner merge of chain and wheel. Both set the big poster correctly. Qwen-Image is cheaper and faster: $0.02 a megapixel against $0.05, read on 10 September and 22 October 2026, and about 7 seconds a picture against MOSS's 14 to 30. MOSS's layout planner also adds subtitles and sub-copy you did not ask for. One run per prompt is a fair first look, not a ranking.
Side by side, measured
| MOSS-Image | Qwen-Image | |
|---|---|---|
| Vendor | OpenMOSS | Alibaba |
| Released | 2026-10-16 | 2025-08-04 |
| Price | $0.05 per megapixel via fal | $0.02 per megapixel via fal |
| Seconds per image | 14 to 63 | 5 to 26 |
| 12 pictures cost | $0.6979 | $0.38 |
| Sizes | Any width and height from 512 to 2048 pixels, rounded to multiples of 16 | Any size up to about 2K through fal's presets or custom pixels |
Seconds and cost are from our own runs of the same 12 prompts through each model.
The same 12 prompts, MOSS-Image against Qwen-Image
One try each, nothing re-rolled or retouched, generated by us. MOSS-Image is on the left, Qwen-Image on the right. Every prompt has its own page with the whole roster.
Text in an image
Every model on this prompt →Both spelled all three lines. Qwen-Image's is a bold gig poster; MOSS-Image's adds a two-colour sun-and-soundwave panel.
›What we asked for
A concert poster that reads 'MODA SUMMER SESSIONS' in large bold type, with the subtitle 'Friday 12 June, 8pm, Roundhouse' and a smaller line 'Tickets at the door'. Risograph style, two colours, tangerine and navy, grainy ink texture.
Product on white
Every model on this prompt →
›What we asked for
Studio product photograph of a matte black ceramic pour-over coffee dripper on a seamless white background, soft top light, subtle contact shadow, centred, catalogue style.
Face and hands
Every model on this prompt →
›What we asked for
Close-up portrait of a woman in her 40s with short grey hair, laughing, holding a red ceramic mug with both hands, natural window light, 85mm lens, shallow depth of field.
Logo mark
Every model on this prompt →Both combined a chain and a wheel. MOSS-Image merged them into one mark; Qwen-Image set them side by side with a spike.
›What we asked for
A flat vector logo mark for a bicycle repair shop called 'Spoke & Chain': a single-colour icon combining a chain link and a wheel spoke, with the wordmark below in a geometric sans-serif, black on white.
Busy scene
Every model on this prompt →
›What we asked for
A crowded Saturday farmers market seen from slightly above: at least twelve stalls with striped awnings, shoppers carrying tote bags, a dog on a leash, crates of oranges, a musician with a guitar, bunting overhead, late morning light.
Specific art style
Every model on this prompt →
›What we asked for
A lighthouse on a rocky cliff during a storm, in the style of a Japanese ukiyo-e woodblock print, with visible wood grain texture, flat colour areas and a bokashi gradient sky.
MOSS-Image drew six bars to their values on a correct axis. Qwen-Image's month labels are garbled.
›What we asked for
A clean bar chart titled 'Monthly active users' showing six months, January through June, with the values 120, 150, 180, 170, 220 and 260, labelled axes, white background, one accent colour, no 3D effects.
Wide aspect ratio
Every model on this prompt →
›What we asked for
Website hero banner: a wide panorama of a sunlit modern co-working space with plants, the left third of the frame kept as clear wall space for text, warm neutral palette.
Infographic with steps
Every model on this prompt →MOSS-Image kept every step exact and added sub-copy. Qwen-Image misspelled the title and lost step 3's wording.
›What we asked for
An infographic card titled 'How to brew pour-over' with four numbered steps in this order: 1. Rinse the filter, 2. Bloom for 30 seconds, 3. Pour in slow circles, 4. Wait 3 minutes. A simple icon beside each step, muted palette, clean sans-serif type.
Food photography
Every model on this prompt →
›What we asked for
Overhead photograph of a margherita pizza on a wooden board, fresh basil leaves, one slice pulled out with the cheese stretching, restaurant menu photography, natural light.
Hard composition
Every model on this prompt →
›What we asked for
A glass of water on a wooden table with exactly three ice cubes in it, a yellow pencil resting across the rim of the glass, a blue book to the left of the glass and a red apple to the right, photographed from table height.
Edit an existing image
Every model on this prompt →›What we asked for
Change the coffee dripper to a deep forest green and add a small brass handle on its right side. Keep the background, lighting and composition exactly as they are.
Frequently asked questions
Which open model is better at text?
On our twelve prompts, MOSS-Image: its chart and infographic were correct where Qwen-Image garbled the small labels. Both set a large three-line poster correctly.
Which is cheaper?
Qwen-Image, at $0.02 a megapixel on fal against $0.05 for MOSS-Image.
Can I run both myself?
Yes. Qwen-Image's weights are Apache 2.0; MOSS-Image's pipeline is Apache 2.0 and its checkpoints are published by OpenMOSS.
Voyager
Run any of these inside Voyager
Voyager is an agent for creative work that runs every model on this page. Brief it, and it picks the model for the job, makes the thing and hands back something you can edit. Create an account to get started.