Disclosure: This article was written with AI assistance and reviewed by EmpirioLabs AI.
Short answer: on three prompts that need lettering, MOSS-Image Turbo, Qwen Image 2.1 (Turbo tier), and Seedream 5.0 Flash spelled every requested word correctly, in English and Chinese, at $0.015, $0.016, and $0.030 an image. FLUX.2 Klein 4B is the lowest price at $0.0085 and misspelled the text in all three, so use it for images without words and for edits from reference images. All four run on EmpirioLabs through one API.
Specs at a glance
| MOSS-Image Turbo | FLUX.2 Klein 4B | Qwen Image 2.1 (Turbo) | Seedream 5.0 Flash | |
|---|---|---|---|---|
| Price per image | $0.015 | $0.0085 | $0.016 | $0.030 |
| Requested text correct (of 3) | 3 | 0 | 3 | 3 |
| Image inputs and editing | No | Yes | Yes | Yes |
| Largest output | 2,048 px on a side, 1 megapixel | 2,048 px on a side, 1 megapixel | 2K presets or an exact size | Up to 2K tiers or an exact size |
How we ran it
Each model got the same three prompts, one shot each, with the same seed (20261024) and its default settings at 1024x1024 (the 1K tier at 1:1 for Qwen Image 2.1 and Seedream 5.0 Flash), through POST /v1/images/generations on October 24, 2026. No result was retried or picked.
Prompt 1: an English trail sign
A hand-painted wooden sign at a mountain trailhead that reads "SUMMIT 2.4 MILES" with a small arrow, pine forest in soft morning fog, photorealistic
Prompt 2: a jazz poster
A minimalist poster for a jazz night with the title "BLUE NOTE SESSIONS" in large white serif letters and the subtitle "Friday 9 PM, Free Entry", a saxophone silhouette on deep navy
Prompt 3: a Chinese shop sign
一家茶叶店的木质招牌，上面用金色楷书写着“春山茶舍”四个字，旁边挂着红灯笼，傍晚暖色灯光，真实摄影 (a tea shop's wooden sign with the four characters 春山茶舍 in gold regular script, red lanterns, warm evening light)
When to use each
- MOSS-Image Turbo for signs, posters, and labels in English or Chinese from text alone, at a low price per image.
- Qwen Image 2.1 when the same request also needs editing or several reference images; its Pro tier adds quality at a higher price.
- Seedream 5.0 Flash for larger outputs, layer decomposition, and interactive edits.
- FLUX.2 Klein 4B for images without lettering and for edits from reference images, at the lowest price of the four.
Generate them on EmpirioLabs
curl https://api.empiriolabs.ai/v1/images/generations \
-H "Authorization: Bearer $EMPIRIOLABS_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "moss-image-turbo",
"prompt": "A hand-painted wooden sign that reads \"SUMMIT 2.4 MILES\"",
"size": "1024x1024",
"sync": true
}'Change model to flux-2-klein-4b, qwen-image-2-1, or seedream-5-0-flash to run the same prompt on another model; each model page lists its own size options.
Frequently asked questions
Which model spells text most reliably?
In this test MOSS-Image Turbo, Qwen Image 2.1, and Seedream 5.0 Flash got all three texts right, including the Chinese sign, and FLUX.2 Klein 4B got none right.
Can MOSS-Image Turbo edit an image?
No. It is text to image only and refuses image inputs; Qwen Image 2.1, Seedream 5.0 Flash, and FLUX.2 Klein 4B accept reference images.
How large can the images be?
MOSS-Image Turbo and FLUX.2 Klein 4B go up to 2,048 pixels on a side and 1 megapixel in total; Qwen Image 2.1 and Seedream 5.0 Flash offer 2K outputs.
How do I get a result without polling?
Send "sync": true and the response is the finished image.
Try it
Playground | Models | Pricing
