Unprompted

Operated by Skald Studio, which sells AI visibility work. No placement on this chart is for sale.

AI Image Generators · measured 2026-11-16 · reading 2026-11-17

Check the answers behind the chart.

Read the original AI answers behind each brand mention. Each question is scheduled for 5 attempts per assistant. Errors and refusals are shown but excluded from mention percentages.

Questions in reading

15

The question list for this reading.

Answers

450

Of 450 attempts in the full reading, before filters.

Errors or refusals

0

Excluded from brand mention percentages.

Answer coverage by assistant

The share of recorded attempts that produced an answer, before filters.

ChatGPT75/75
Claude75/75
Claude Code (local)75/75
Codex (local)75/75
Gemini75/75
Perplexity75/75

Recorded answers

Each line lists the brands found in one answer.

Choose a question to load its original answers and recorded source links. Sources may be citations or search results; a link alone does not show which brand it supports.

q01

What is the best AI image generator? #

Named first most often: Midjourney (15 of 30 answers with these filters).

30 recorded attempts with these filters.

Read full answers to this question

ChatGPT 5 recorded answers
DALL-E · Midjourney · Ideogram · Adobe Firefly
DALL-E · Midjourney · Ideogram · Adobe Firefly
Midjourney · Ideogram · Adobe Firefly
DALL-E · Midjourney · Imagen · Adobe Firefly
DALL-E · Midjourney · Ideogram
Claude 5 recorded answers
Imagen · Midjourney · Ideogram · Adobe Firefly · Recraft · Flux · Stable Diffusion
Imagen · Midjourney · Ideogram · Adobe Firefly
Midjourney · Ideogram · Recraft · Adobe Firefly · Flux
Imagen · Midjourney · Ideogram · Recraft · Adobe Firefly · Flux
Imagen · Midjourney · Adobe Firefly · Ideogram · Flux
Claude Code (local) 5 recorded answers
DALL-E · Midjourney · Flux · Ideogram · Adobe Firefly
Imagen · Flux · Midjourney · Adobe Firefly · Ideogram
Midjourney · Imagen · Flux · Ideogram · Adobe Firefly
Midjourney · Imagen · Adobe Firefly · Ideogram
Midjourney · Imagen · Adobe Firefly · Ideogram · Flux · Stable Diffusion
Codex (local) 5 recorded answers
Midjourney · Adobe Firefly · Ideogram
Midjourney · Adobe Firefly
Midjourney · Ideogram · Adobe Firefly
Midjourney · Adobe Firefly
Midjourney · Adobe Firefly · Ideogram
Gemini 5 recorded answers
Midjourney · Flux · NightCafe · Ideogram · Adobe Firefly
Midjourney · Flux · Ideogram · Adobe Firefly · Stable Diffusion
Midjourney · Ideogram · Recraft · Adobe Firefly · Flux · Stable Diffusion · Higgsfield
Midjourney · DALL-E · Imagen · Flux · NightCafe · Adobe Firefly
Midjourney · Flux · NightCafe · Adobe Firefly · Recraft · Ideogram
Perplexity 5 recorded answers
DALL-E · Midjourney · Ideogram · Adobe Firefly
Adobe Firefly · Midjourney · DALL-E
Imagen · DALL-E · Midjourney · Adobe Firefly · Ideogram
Imagen · DALL-E · Midjourney · Adobe Firefly · Recraft
Imagen · Adobe Firefly · Midjourney · DALL-E

q02

Which AI makes the most realistic images? #

Named first most often: Flux (11 of 30 answers with these filters).

30 recorded attempts with these filters.

Read full answers to this question

ChatGPT 5 recorded answers
Imagen
DALL-E · Midjourney · Imagen
DALL-E · Imagen
DALL-E · Midjourney
DALL-E · Midjourney
Claude 5 recorded answers
Flux · Adobe Firefly · Midjourney
Flux · Midjourney
Imagen · Midjourney · Flux · Adobe Firefly
Flux · Midjourney · Ideogram
Flux · Midjourney · Ideogram
Claude Code (local) 5 recorded answers
Imagen · DALL-E · Flux · Krea · Freepik · getimg.ai · Midjourney · Seedream
Imagen · DALL-E · Flux · Krea · Freepik · Midjourney · Seedream
DALL-E · Imagen · Flux · Krea · Freepik · Leonardo AI · Seedream · Midjourney
Imagen · Flux · Krea · Freepik · DALL-E · Midjourney · Seedream
Imagen · DALL-E · Flux · Krea · Freepik · getimg.ai · Midjourney · Seedream
Codex (local) 5 recorded answers
No recognized brand was found
No recognized brand was found
Imagen · Midjourney
DALL-E · Midjourney · Flux
No recognized brand was found
Gemini 5 recorded answers
Flux · NightCafe · Midjourney · Imagen
Flux · NightCafe · Midjourney · Imagen · Leonardo AI
Flux · NightCafe · Freepik · Midjourney · Imagen · Stable Diffusion
Flux · Freepik · Midjourney · Imagen · Leonardo AI
Flux · Freepik · Midjourney · Imagen
Perplexity 5 recorded answers
Flux · Midjourney
Imagen · Flux · DALL-E · Midjourney
DALL-E · Imagen · Midjourney
Flux · Imagen · Midjourney
Imagen · Flux

q03

What is the best AI image generator for beginners? #

Named first most often: Canva (12 of 30 answers with these filters).

30 recorded attempts with these filters.

Read full answers to this question

ChatGPT 5 recorded answers
Canva
Canva
Canva
DALL-E · Canva
DALL-E · Canva
Claude 5 recorded answers
Imagen · Canva · Bing Image Creator · Adobe Firefly · Midjourney
Imagen · Bing Image Creator · Canva · Adobe Firefly · Leonardo AI · Midjourney
Imagen · Canva · Adobe Firefly · Midjourney
Imagen · Canva · Adobe Firefly · Midjourney
Imagen · Canva · Adobe Firefly · Midjourney
Claude Code (local) 5 recorded answers
Imagen · Canva · Adobe Firefly · Bing Image Creator · Midjourney
Imagen · Canva · Adobe Firefly · Midjourney
Imagen · Bing Image Creator · Canva · Adobe Firefly · Midjourney · Ideogram
Imagen · Bing Image Creator · Canva · Adobe Firefly · Midjourney · Ideogram
Imagen · Canva · Adobe Firefly · Midjourney
Codex (local) 5 recorded answers
Canva
Midjourney
Canva
Canva
Canva · Adobe Firefly
Gemini 5 recorded answers
DALL-E · Adobe Firefly · Canva · Ideogram
Imagen · Adobe Firefly · Canva · Ideogram
Adobe Firefly · Canva
Adobe Firefly · Canva
Bing Image Creator · Adobe Firefly · Canva · Midjourney
Perplexity 5 recorded answers
Canva · DALL-E · Adobe Firefly
Canva · Adobe Firefly
Canva · Adobe Firefly
Canva · Adobe Firefly
Canva

q04

Which AI image tool is best for commercial use? #

Named first most often: Adobe Firefly (30 of 30 answers with these filters).

30 recorded attempts with these filters.

Read full answers to this question

ChatGPT 5 recorded answers
Adobe Firefly · Midjourney
Adobe Firefly · Midjourney · DALL-E
Adobe Firefly · Midjourney
Adobe Firefly · Midjourney
Adobe Firefly · Midjourney
Claude 5 recorded answers
Adobe Firefly · Midjourney · Ideogram · Flux
Adobe Firefly · Midjourney · Ideogram · Leonardo AI
Adobe Firefly · Midjourney · Ideogram · Flux
Adobe Firefly · Midjourney
Adobe Firefly · Midjourney · Ideogram · Flux · Stable Diffusion
Claude Code (local) 5 recorded answers
Adobe Firefly · Midjourney · Flux
Adobe Firefly · Midjourney · Ideogram · Stable Diffusion · Flux
Adobe Firefly · Midjourney · Ideogram · Stable Diffusion · Flux
Adobe Firefly · Midjourney · Imagen · Stable Diffusion
Adobe Firefly · Midjourney · Imagen · Ideogram · Stable Diffusion · Flux
Codex (local) 5 recorded answers
Adobe Firefly · Midjourney
Adobe Firefly · Midjourney
Adobe Firefly · Midjourney
Adobe Firefly · DALL-E · Midjourney
Adobe Firefly · Midjourney
Gemini 5 recorded answers
Adobe Firefly · Midjourney · Ideogram · Flux · Leonardo AI
Adobe Firefly · Midjourney · Flux · Freepik · NightCafe · Canva · Leonardo AI
Adobe Firefly · Midjourney · Ideogram · Flux · Stable Diffusion
Adobe Firefly · Midjourney · Ideogram · Flux
Adobe Firefly · Midjourney · Canva
Perplexity 5 recorded answers
Adobe Firefly · Midjourney
Adobe Firefly · Midjourney
Adobe Firefly · Midjourney
Adobe Firefly · Midjourney
Adobe Firefly · Midjourney

q05

What is the best free AI image generator? #

Named first most often: Imagen (16 of 30 answers with these filters).

30 recorded attempts with these filters.

Read full answers to this question

ChatGPT 5 recorded answers
Imagen · Bing Image Creator
Bing Image Creator · Imagen
Leonardo AI · Ideogram
Imagen · Bing Image Creator · Ideogram
Imagen · Ideogram · Bing Image Creator
Claude 5 recorded answers
Imagen · Bing Image Creator · Ideogram · Adobe Firefly · Canva · Leonardo AI · Stable Diffusion · Flux
Imagen · Bing Image Creator · Ideogram · Leonardo AI · Adobe Firefly
Imagen · Bing Image Creator · Ideogram · Adobe Firefly · Leonardo AI · Krea · Canva · Stable Diffusion · Flux
Imagen · Bing Image Creator · Ideogram · Adobe Firefly · Leonardo AI · Canva
Imagen · Bing Image Creator · Adobe Firefly · Ideogram · Krea · Midjourney · Recraft
Claude Code (local) 5 recorded answers
Bing Image Creator · Imagen · Ideogram · Leonardo AI · Stable Diffusion · Flux
Imagen · Bing Image Creator · Stable Diffusion · Flux · Ideogram · Leonardo AI · Adobe Firefly
Bing Image Creator · Imagen · Ideogram · Stable Diffusion · Flux · Leonardo AI · Adobe Firefly
Imagen · Bing Image Creator · Ideogram · Leonardo AI · Adobe Firefly · Stable Diffusion · Flux
Imagen · Bing Image Creator · Ideogram · Leonardo AI · Adobe Firefly · Stable Diffusion · Flux
Codex (local) 5 recorded answers
Bing Image Creator · Adobe Firefly
Bing Image Creator · Ideogram
Bing Image Creator · Ideogram
Bing Image Creator · Adobe Firefly
Bing Image Creator · Ideogram · Adobe Firefly
Gemini 5 recorded answers
Bing Image Creator · Ideogram · Leonardo AI · Imagen · Adobe Firefly · Krea · Flux
Bing Image Creator · Ideogram · Leonardo AI · Adobe Firefly
Bing Image Creator · Ideogram · Leonardo AI · Flux
Ideogram · Leonardo AI · Bing Image Creator · Adobe Firefly
Bing Image Creator · Ideogram · Leonardo AI · Krea · Imagen
Perplexity 5 recorded answers
Imagen · Canva · Leonardo AI · Ideogram
Imagen · Ideogram
Imagen · Ideogram · Bing Image Creator
Imagen · Bing Image Creator
Imagen · Bing Image Creator · Ideogram

q06

Which AI is best at generating text inside images? #

Named first most often: Ideogram (20 of 30 answers with these filters).

30 recorded attempts with these filters.

Read full answers to this question

ChatGPT 5 recorded answers
DALL-E · Ideogram
DALL-E · Ideogram · Imagen
DALL-E · Ideogram
DALL-E · Ideogram
DALL-E · Ideogram
Claude 5 recorded answers
Ideogram · Imagen
Ideogram · DALL-E · Imagen
Ideogram · Flux · Qwen-Image
Imagen · Ideogram
Ideogram · DALL-E · Imagen
Claude Code (local) 5 recorded answers
Ideogram · Midjourney · Flux · Adobe Firefly
DALL-E · Ideogram · Midjourney · Flux
DALL-E · Ideogram · Canva
Ideogram · DALL-E · Midjourney · Flux · Canva
Ideogram · Midjourney · Flux
Codex (local) 5 recorded answers
Ideogram · Canva
Ideogram · Canva
Ideogram · Canva
Ideogram · Canva
Ideogram
Gemini 5 recorded answers
Ideogram · Flux · Freepik · Recraft · DALL-E · Canva
Ideogram · Flux · NightCafe · Recraft
Ideogram · Flux · Canva
Ideogram · Recraft · Flux · DALL-E
Ideogram · Flux · Midjourney · DALL-E
Perplexity 5 recorded answers
Ideogram · Imagen
Ideogram · DALL-E · Imagen · Canva
Ideogram · DALL-E · Imagen
DALL-E · Ideogram
DALL-E · Ideogram

q07

What AI image generator do professional designers use? #

Named first most often: Adobe Firefly (20 of 30 answers with these filters).

30 recorded attempts with these filters.

Read full answers to this question

ChatGPT 5 recorded answers
Midjourney · Adobe Firefly · Recraft · Ideogram
Midjourney · Adobe Firefly · DALL-E
Midjourney · Adobe Firefly
Midjourney · Adobe Firefly · DALL-E
Midjourney · Adobe Firefly · DALL-E
Claude 5 recorded answers
Adobe Firefly · Midjourney · Ideogram · Flux · Stable Diffusion · Canva
Adobe Firefly · Midjourney · Imagen · Ideogram · Recraft · Flux · Stable Diffusion · Canva
Adobe Firefly · Midjourney · Ideogram · Recraft · Stable Diffusion · Flux
Adobe Firefly · Midjourney · Ideogram · Stable Diffusion · Flux · Canva
Adobe Firefly · Midjourney · DALL-E · Imagen · Ideogram · Recraft · Krea · Flux · Stable Diffusion
Claude Code (local) 5 recorded answers
Adobe Firefly · Midjourney · Ideogram · Recraft · Flux · Krea · Freepik
Adobe Firefly · Midjourney · Ideogram · Recraft · Flux · Stable Diffusion
Adobe Firefly · Midjourney · Ideogram · Stable Diffusion · Flux · Krea · Freepik
Adobe Firefly · Midjourney · Imagen · Ideogram · Stable Diffusion · Flux · Canva · Krea
Adobe Firefly · Midjourney · Stable Diffusion · Flux · Ideogram · Krea · Leonardo AI
Codex (local) 5 recorded answers
Adobe Firefly · Midjourney · DALL-E
Midjourney · Adobe Firefly · DALL-E · Ideogram
Midjourney · Adobe Firefly · DALL-E
Midjourney · Adobe Firefly · DALL-E
Adobe Firefly · Midjourney · DALL-E
Gemini 5 recorded answers
Adobe Firefly · Midjourney · Recraft · Ideogram · Stable Diffusion · Flux
Adobe Firefly · Midjourney · Recraft · Flux · Ideogram · Stable Diffusion
Adobe Firefly · Midjourney · Recraft · Flux · Stable Diffusion · Ideogram · DALL-E
Adobe Firefly · Midjourney · Recraft · Ideogram · Flux · Stable Diffusion
Adobe Firefly · Midjourney · Recraft · Flux · Stable Diffusion · Ideogram
Perplexity 5 recorded answers
Adobe Firefly · Midjourney
Adobe Firefly · Midjourney · Ideogram · Recraft
Midjourney · Adobe Firefly
Midjourney · Adobe Firefly · Ideogram
Adobe Firefly · Midjourney · Ideogram

q08

Which AI image generator gives the best value for money? #

Named first most often: Imagen (10 of 30 answers with these filters).

30 recorded attempts with these filters.

Read full answers to this question

ChatGPT 5 recorded answers
Leonardo AI · Midjourney
Adobe Firefly
Leonardo AI · Midjourney
Adobe Firefly · Midjourney
Adobe Firefly · Leonardo AI
Claude 5 recorded answers
Imagen · Midjourney · Ideogram · Bing Image Creator · Adobe Firefly · Stable Diffusion · Flux
Imagen · Midjourney · Leonardo AI · Ideogram · Adobe Firefly · Stable Diffusion · Flux
Imagen · Midjourney · Ideogram · Leonardo AI · Adobe Firefly · Stable Diffusion · Flux
Midjourney · Adobe Firefly · Stable Diffusion · Flux · Ideogram
Imagen · Bing Image Creator · Midjourney · Ideogram · Leonardo AI · Stable Diffusion · Flux
Claude Code (local) 5 recorded answers
Imagen · Midjourney · Bing Image Creator · Leonardo AI · Ideogram · Stable Diffusion · Flux
Imagen · Bing Image Creator · Ideogram · Leonardo AI · Midjourney · Stable Diffusion · Flux
Midjourney · Leonardo AI · Flux · Stable Diffusion
Midjourney · Imagen · Bing Image Creator · Ideogram · Adobe Firefly · Stable Diffusion · Flux
Imagen · Leonardo AI · Midjourney · Ideogram · Stable Diffusion · Flux
Codex (local) 5 recorded answers
Ideogram · Midjourney
Ideogram · Midjourney
Ideogram · Midjourney
Ideogram · Midjourney
Ideogram · Midjourney
Gemini 5 recorded answers
Leonardo AI · Bing Image Creator · Midjourney · Ideogram · Flux · Stable Diffusion
Midjourney · Bing Image Creator · Leonardo AI · Ideogram · Flux
Leonardo AI · Midjourney · Bing Image Creator
Flux · Stable Diffusion · Imagen · Leonardo AI · Midjourney · Ideogram
Flux · Stable Diffusion · Midjourney · Leonardo AI
Perplexity 5 recorded answers
Imagen · Adobe Firefly · Midjourney
Imagen · Midjourney
Leonardo AI · Midjourney
Midjourney
Imagen · Midjourney

q09

What is the best AI for generating logos? #

Named first most often: Recraft (14 of 30 answers with these filters).

30 recorded attempts with these filters.

Read full answers to this question

ChatGPT 5 recorded answers
No recognized brand was found
Canva · Recraft
Canva · Recraft
Recraft · Canva
Recraft · Canva
Claude 5 recorded answers
Recraft · Ideogram · Canva
Recraft · Ideogram · Canva
Recraft · Ideogram · Adobe Firefly · Canva
Canva · Ideogram · Midjourney · Recraft
Recraft · Ideogram · Canva
Claude Code (local) 5 recorded answers
Recraft · Ideogram · Midjourney · Canva
Ideogram · Recraft · Midjourney · Adobe Firefly · Canva
Ideogram · Recraft · Midjourney
Ideogram · Recraft · Midjourney · Canva · Adobe Firefly
Recraft · Ideogram · Adobe Firefly · Midjourney · Canva
Codex (local) 5 recorded answers
Ideogram
Recraft · Ideogram
Recraft
Recraft · Ideogram
Recraft · Ideogram
Gemini 5 recorded answers
Adobe Firefly · Recraft · Ideogram · Canva · Midjourney
Canva · Ideogram
Recraft · Ideogram · Canva
Canva · Recraft · Ideogram · Midjourney
Recraft · Ideogram · Midjourney · Canva
Perplexity 5 recorded answers
Canva
Canva · Recraft
Canva · Recraft
Canva
Canva · Recraft

q10

Which AI image generator has the best style control? #

Named first most often: Midjourney (23 of 30 answers with these filters).

30 recorded attempts with these filters.

Read full answers to this question

ChatGPT 5 recorded answers
Midjourney · Adobe Firefly
Midjourney · Adobe Firefly · Ideogram
Midjourney · Adobe Firefly
Midjourney
Midjourney
Claude 5 recorded answers
Midjourney · Stable Diffusion · Flux · Leonardo AI · getimg.ai · Ideogram
Midjourney · Stable Diffusion · Flux · Recraft · Leonardo AI · Ideogram
Midjourney · Stable Diffusion · Flux · Leonardo AI · Ideogram
Midjourney · Stable Diffusion · Flux · Leonardo AI · Ideogram
Midjourney · Stable Diffusion · Flux · Ideogram · Adobe Firefly
Claude Code (local) 5 recorded answers
Midjourney · Stable Diffusion · Flux · Krea · Adobe Firefly · Leonardo AI
Midjourney · Stable Diffusion · Flux · Krea · Adobe Firefly · Leonardo AI · Ideogram
Midjourney · Stable Diffusion · Flux · Adobe Firefly · Leonardo AI · Imagen
Midjourney · Stable Diffusion · Flux · Leonardo AI · Adobe Firefly · Ideogram · Imagen
Midjourney · Stable Diffusion · Flux · Adobe Firefly · Leonardo AI · Recraft · Imagen
Codex (local) 5 recorded answers
Midjourney · Adobe Firefly
Midjourney · Adobe Firefly
Midjourney · Adobe Firefly
Midjourney · Recraft · Adobe Firefly
Midjourney · Adobe Firefly
Gemini 5 recorded answers
Flux · Stable Diffusion · Midjourney · Leonardo AI · Recraft · Adobe Firefly
Stable Diffusion · Flux · Leonardo AI · Midjourney · Recraft
Stable Diffusion · Flux · Midjourney · Recraft · Adobe Firefly · Leonardo AI
Recraft · Midjourney · Adobe Firefly · Stable Diffusion · Flux
Stable Diffusion · Flux · Recraft · Leonardo AI · Midjourney
Perplexity 5 recorded answers
Midjourney · Adobe Firefly
Adobe Firefly · Midjourney
Midjourney · Adobe Firefly
Adobe Firefly
Midjourney · Adobe Firefly

q11

What is the best AI image generator for marketing images? #

Named first most often: Adobe Firefly (18 of 30 answers with these filters).

30 recorded attempts with these filters.

Read full answers to this question

ChatGPT 5 recorded answers
Adobe Firefly · DALL-E · Ideogram · Canva
Adobe Firefly · Canva
Adobe Firefly · Canva
Adobe Firefly · Canva · DALL-E
Adobe Firefly · Canva · DALL-E
Claude 5 recorded answers
Adobe Firefly · Canva · Ideogram · Recraft · Midjourney · Imagen
DALL-E · Imagen · Midjourney · Canva · Adobe Firefly · Ideogram · Recraft
Midjourney · Adobe Firefly · Canva · Ideogram · Imagen · Recraft
Imagen · Midjourney · Adobe Firefly · Ideogram · Recraft · Canva
Imagen · Midjourney · Ideogram · Adobe Firefly · Canva · Recraft
Claude Code (local) 5 recorded answers
Midjourney · Adobe Firefly · Ideogram · Canva · Imagen · Flux
Adobe Firefly · Midjourney · Ideogram · Canva · Flux · Freepik · Krea · Leonardo AI
Adobe Firefly · Midjourney · Ideogram · Canva · Leonardo AI · getimg.ai
Adobe Firefly · Midjourney · Canva
Adobe Firefly · Midjourney · Canva · Ideogram
Codex (local) 5 recorded answers
Adobe Firefly · DALL-E · Midjourney
Adobe Firefly · DALL-E · Canva · Midjourney
Adobe Firefly · DALL-E · Canva
DALL-E · Canva · Adobe Firefly
DALL-E · Adobe Firefly · Canva
Gemini 5 recorded answers
Midjourney · Ideogram · Adobe Firefly · Canva
Adobe Firefly · Midjourney · Ideogram · Canva
Midjourney · Adobe Firefly · Ideogram · Canva · Recraft · getimg.ai
Adobe Firefly · Midjourney · Ideogram · Canva
Adobe Firefly · Ideogram · Midjourney · Canva · Recraft
Perplexity 5 recorded answers
Canva · Adobe Firefly
Adobe Firefly · Canva
Canva · Adobe Firefly · Midjourney · Ideogram
Adobe Firefly · Canva · Midjourney · Ideogram
Canva · Adobe Firefly

q12

Which AI image tool is best for editing existing photos? #

Named first most often: Imagen (13 of 30 answers with these filters).

30 recorded attempts with these filters.

Read full answers to this question

ChatGPT 5 recorded answers
DALL-E
Imagen
DALL-E
DALL-E
DALL-E
Claude 5 recorded answers
Imagen · Flux
Imagen · Flux
Imagen · Canva
Flux · Canva
Imagen · Flux · Canva
Claude Code (local) 5 recorded answers
Imagen
Imagen · Canva
Imagen · Canva · Flux · Stable Diffusion
Imagen
Imagen · Canva · Flux · Krea · Freepik · Stable Diffusion
Codex (local) 5 recorded answers
Imagen · DALL-E
No recognized brand was found
DALL-E
Canva
No recognized brand was found
Gemini 5 recorded answers
Adobe Firefly · Canva
Canva
Canva
Adobe Firefly · Canva
Canva
Perplexity 5 recorded answers
Imagen
Imagen
Canva
Canva
Canva

q13

What is the best open source AI image generator? #

Named first most often: Flux (25 of 30 answers with these filters).

30 recorded attempts with these filters.

Hide full answer text

ChatGPT 5 recorded answers
Flux · Qwen-Image
Read answer 1 · 18:39 UTC

**My pick is FLUX.2 [klein] 4B, run locally in ComfyUI.** It’s a strong all-around choice for generating and editing images, and its Apache 2.0 license permits commercial use. ComfyUI is the open-source interface; FLUX is the model that makes the images. Plan on roughly **13 GB of GPU memory** for a straightforward local setup. ([help.bfl.ai](https://help.bfl.ai/articles/6122710168-which-flux-2-model-should-i-choose?utm_source=openai)) If your images need **accurate words on signs or posters**, also try **Qwen-Image-2512**. It’s Apache 2.0-licensed and emphasizes text rendering. ([huggingface.co](https://huggingface.co/Qwen/Qwen-Image-2512/blob/main/README.md?utm_source=openai)) One licensing caution: *open weights* doesn’t always mean unrestricted open source. The newer **FLUX 3 Image Klein** has different commercial terms, so I wouldn’t substitute it for FLUX.2 [klein] 4B without checking its license. ([bfl.ai](https://bfl.ai/models/flux-3-image-klein?utm_source=openai))

Sources were recorded as citations in the answer.

Qwen-Image · Flux · Ideogram
Read answer 2 · 18:39 UTC

**My pick is Qwen-Image-2512** if you mean a model you can download and use under a permissive license. It’s strong at realistic images and putting readable text in images, and its weights are released under Apache 2.0. Run it through **ComfyUI** if you want a visual interface rather than writing code. ([huggingface.co](https://huggingface.co/Qwen/Qwen-Image-2512/blob/main/README.md?utm_source=openai)) If you have a more modest GPU, try **FLUX.2 [klein] 4B** instead. It handles both generation and editing, is Apache 2.0–licensed, and its developer says it runs on GPUs with about 13 GB of VRAM. ([huggingface.co](https://huggingface.co/black-forest-labs/FLUX.2-klein-4B?hardware=mi250&utm_source=openai)) One licensing trap: **Ideogram 4.0** is an impressive *open-weight* option, but its freely downloadable weights are **non-commercial**, not permissively open-source. ([ideogram.ai](https://ideogram.ai/licensing/?utm_source=openai))

Sources were recorded as citations in the answer.

Flux · Ideogram
Read answer 3 · 18:39 UTC

**My pick: ComfyUI running FLUX.2 [klein] 4B.** ComfyUI is an open-source app you can run locally, and FLUX.2 [klein] 4B generates and edits images with Apache 2.0–licensed weights, including commercial use. Plan on roughly **13 GB of GPU memory** for the model. ([github.com](https://github.com/comfy-org/ComfyUI?utm_source=openai)) If by “best” you mean **image quality above licensing flexibility**, Ideogram 4.0 is also worth a look—but its downloadable weights are *not* freely licensed for commercial deployment. ([ideogram.ai](https://ideogram.ai/blog/ideogram-4.0/?utm_source=openai))

Sources were recorded as citations in the answer.

Z-Image · Flux
Read answer 4 · 18:39 UTC

**My pick is Z-Image-Turbo** if you want an open-source model for generating images locally. It’s fast, runs on consumer hardware, and its Apache 2.0 license permits commercial use. You can run it through **ComfyUI**. ([arxiv.org](https://arxiv.org/abs/2511.22699?utm_source=openai)) If you also want to **edit existing images**, choose **FLUX.2 [klein] 4B** instead. It combines generation and editing, runs on a GPU with roughly 13 GB of VRAM, and is also Apache 2.0–licensed. ([huggingface.co](https://huggingface.co/black-forest-labs/FLUX.2-klein-4B?hardware=mi250&utm_source=openai))

Sources were recorded as citations in the answer.

Flux · Qwen-Image
Read answer 5 · 18:39 UTC

**My pick for most people is FLUX.2 [klein] 4B.** It generates and edits images, runs locally on a consumer GPU with roughly 13 GB of VRAM, works with ComfyUI, and is released under the commercially usable Apache 2.0 license. It’s a strong balance of quality, speed, and practicality. ([huggingface.co](https://huggingface.co/black-forest-labs/FLUX.2-klein-4B)) If **image quality and text in images** matter more than running a compact model, try **Qwen-Image-2512** instead. It’s also Apache 2.0, but it’s a heavier setup. ([huggingface.co](https://huggingface.co/Qwen/Qwen-Image-2512)) A small distinction: these are *open-weight* models with permissive licenses; that doesn’t necessarily mean their full training data and process are open.

Sources were recorded as citations in the answer.

Claude 5 recorded answers
Qwen-Image · Flux · Z-Image · Stable Diffusion
Read answer 1 · 18:28 UTC

There isn't one clear winner. The leaders change every few months, and most of these models are "open weights" (you can download them) with licenses that may limit commercial use. Here are the best picks right now: **Best image quality: Qwen-Image-2.1 (Alibaba) or FLUX (Black Forest Labs)** - On Artificial Analysis's open-weights text-to-image leaderboard, Qwen-Image-2.1 and the brand-new FLUX 3 Image Klein are close together near the top. - Qwen-Image-2.1 is a unified text-to-image and image-editing model released on September 20, 2026 . Its license is Qwen Research (non-commercial) . - FLUX.2 [dev] is the older, well-tested choice. It's a 32B model that does text-to-image and single/multi-reference editing in one checkpoint , but it's released under the FLUX [dev] Non-Commercial License . **Best for commercial use: Qwen-Image-2512 or Z-Image Turbo** - Qwen-Image-2512 ranks well and is licensed Apache 2.0 . - Z-Image Turbo is a 6B parameter model built by Alibaba's Tongyi-MAI team for fast generation on consumer and enterprise GPUs . It's also Apache 2.0 and needs about 16GB of VRAM. **Best for text inside images (posters, logos, signs): Qwen-Image** - Qwen-Image is the strongest open-source option for text accuracy inside images, with native support for English and Chinese typography. **Best for beginners, fine-tuning and community add-ons: Stable Diffusion (SDXL / SD 3.5)** - Its scores are lower now, but Stable Diffusion built the open-source AI image generator category and still has the deepest ecosystem around it. That means many tutorials, LoRAs (small add-on models for specific styles or characters) and tools. **My recommendation:** - **Personal use with a strong GPU (24GB+):** Qwen-Image-2.1 or FLUX.2 [dev]. - **Business use:** Qwen-Image-2512. - **A mid-range graphics card:** Z-Image Turbo. Whichever you pick, run it through a free front end like ComfyUI.

Sources were recorded as search results.

Qwen-Image · Flux · Z-Image · Stable Diffusion
Read answer 2 · 18:28 UTC

There isn't one clear winner, because the top spot keeps changing. Here's where things stand now. **Highest rated for image quality right now: MiniMax H3-Image.** It's #1 on Artificial Analysis's open-weights leaderboard, which ranks models by human preference votes. H3-Image scores 1066 Elo (Nov 2026), followed by Qwen-Image-2.1 at 1047, FLUX 3 Image Klein at 1041, and Bria's Fibo 2.0 at 1037. The models from #2 to #4 are close enough that they're basically tied. **My practical pick: Black Forest Labs' FLUX family (FLUX 3 Klein or FLUX.2).** It has the biggest community, the most tools and add-ons, and steady quality. FLUX.2 was released by Black Forest Labs on November 25, 2025 , and it can keep a character or style consistent using up to 10 reference images in a single generation. One catch: FLUX.2 [dev] is released under the FLUX [dev] Non-Commercial License . **Check the license before you pick.** Most of these are "open weights" (you can download the model) rather than truly open source, and some ban commercial use. - **Qwen-Image-2.1 (Alibaba):** It tops many rankings, but its license allows research and evaluation only. - **Qwen-Image-2512:** This is the safer choice for commercial work because it's Apache 2.0 . **Other options for specific needs:** - **Speed or a home GPU: Z-Image Turbo.** It's a 6B parameter model from Alibaba's Tongyi-MAI team, built for fast generation on consumer and enterprise GPUs , and it's also Apache 2.0. - **The most fine-tunes, LoRAs and tutorials: Stable Diffusion (SDXL / 3.5).** Pick it if you're new to image generation, want the largest community, or need lots of LoRA fine-tuning with little setup. The trade-off is that its image quality is now well behind the leaders. **Bottom line:** For general use, go with FLUX. If you'll use the images commercially, use Qwen-Image-2512 or Z-Image Turbo. For the highest scores on today's leaderboard, try H3-Image or Qwen-Image-2.1, but read their licenses first. Most people run these models locally in ComfyUI.

Sources were recorded as search results.

Flux · Ideogram · Qwen-Image · Z-Image · Stable Diffusion
Read answer 3 · 18:28 UTC

There isn't one clear winner. The top spot changes every few months, and the best choice depends on whether you care most about quality, commercial use, or your hardware. **Best image quality: Qwen-Image-2.1 (Alibaba) or FLUX.2 [dev] (Black Forest Labs)** - In a ranking dated October 7, 2026, the open-weight leader on the Artificial Analysis Text-to-Image Arena is Alibaba's Qwen-Image-2.1 at 1,035 Elo, released on September 20, with Ideogram 4.0 (1,012) and FLUX.2 [dev] (1,000) next. Black Forest Labs' newer FLUX 3 Image Klein is scoring about the same as Qwen-Image-2.1. - One catch: both models have non-commercial licenses. Qwen-Image-2.1 is listed under a Qwen Research (non-commercial) license, and FLUX.2 [dev] is released under the FLUX [dev] Non-Commercial License . **Best for commercial use: Qwen-Image-2512** - It ranks close to the leaders and is Apache 2.0 (as Qwen-Image-2512) . - It's also especially good at putting readable text in images. Qwen-Image is the strongest open-source option for text accuracy inside images, with native support for English and Chinese typography. **Best for regular gaming GPUs and speed: Z-Image Turbo** - Z-Image-Turbo is a 6B model built for speed, generating images in under a second on enterprise GPUs and running comfortably on a 16GB consumer card. It's also Apache 2.0. **Best ecosystem for beginners: Stable Diffusion (SDXL / 3.5)** - It doesn't produce the best images anymore, but it's still a good starting point. Choose it if you are new to image generation, need access to the largest community ecosystem, or require extensive LoRA fine-tuning with minimal infrastructure overhead. **My recommendation:** If you have a 24GB+ GPU and just want the best-looking images, use Qwen-Image-2.1 or FLUX.2 [dev]. If you'll sell or commercially use what you make, pick Qwen-Image-2512. If you have a 16GB card, start with Z-Image Turbo. Most people run these models through a free app like ComfyUI.

Sources were recorded as search results.

Flux · Stable Diffusion
Read answer 4 · 18:28 UTC

There isn't one best open-source image generator. It depends on your graphics card's memory (VRAM) and whether you'll use the images commercially. These are the models I'd look at right now: **Best overall quality: FLUX.2 by Black Forest Labs.** It came out in November 2025, and current reviews treat it as the quality benchmark for models you can download and run yourself. It can also keep a character or style consistent using up to 10 reference images. The catch is that it needs a powerful GPU, and the [dev] version's license limits commercial use, so read the terms. **Best fully open option for commercial use: Qwen-Image by Alibaba.** It uses the Apache 2.0 license, so commercial use is allowed. It's very good at putting readable text in images, works in multiple languages and also edits images. It needs about 24 GB of VRAM. **Best for speed or ordinary hardware: Z-Image Turbo by Alibaba's Tongyi team.** It's a smaller 6B-parameter model, also Apache 2.0. It makes an image in about 1 second on a data-center H100 card and about 5–10 seconds on an RTX 4090. It needs around 16 GB of VRAM. **Best add-ons and customization: Stable Diffusion 3.5 or SDXL by Stability AI.** The image quality isn't the highest anymore, but they have by far the most community extras: LoRAs (small add-ons that teach a model a specific style or character), fine-tuned versions and tools. They also run on modest GPUs. **Best for photorealism: HunyuanImage 3.0 by Tencent.** It's an 80B-parameter model, which makes it the largest open-source image model, so it really needs data-center hardware. **My pick:** - With a 24 GB+ card, start with FLUX.2 [dev]. If you need simple commercial licensing, use Qwen-Image. - With a 12–16 GB card, use Z-Image Turbo or SDXL. To run any of these, I'd use the free app ComfyUI, which supports almost every model. If you want something simpler, Forge is easier to start with.

Sources were recorded as search results.

Flux · Qwen-Image · Stable Diffusion
Read answer 5 · 18:28 UTC

Black Forest Labs' **FLUX** family generally gives the best image quality you can download and run yourself. If you want a model that's actually open source and allowed for commercial use, **Qwen-Image** from Alibaba is the safer choice. This area changes every few months, so these picks may already be out of date. **My picks:** - **Best overall quality: FLUX.2 [dev].** One 2026 guide calls it the current benchmark for open-weight image quality , and it can keep a character's look and visual style consistent using up to 10 reference images at once . The downsides: it's a 32-billion-parameter model , so you need a powerful graphics card, and it isn't licensed for commercial use. - **Best for consumer GPUs: FLUX.2 [klein] or Qwen-Image Turbo.** Klein comes in 9B and 4B versions . The 9B version beat even the full FLUX.2 [dev] at image editing . In one test on an RTX 3080, FLUX was clearly better at counting objects, holding a style and drawing hands , but Qwen Turbo was faster. Peak VRAM was 7.2 GiB for FLUX and 6.4 GiB for Qwen . That same post also mentions a newer "FLUX 3 Klein," which suggests a newer generation is already out. - **Best for text inside images and commercial use: Qwen-Image.** It's the strongest open-source option for accurate text in images, with native English and Chinese typography . The original Qwen-Image is released under Apache 2.0 . The newer Qwen-Image-2.1 is for non-commercial use only unless you get a separate license . - **Best add-ons and customization: Stable Diffusion 3.5 or SDXL.** It has the largest collection of community add-ons, like style models (LoRAs) and fine-tunes, and runs on as little as 8GB of VRAM . **Licenses:** Many of these are "open weights" rather than truly open source. For example, FLUX.2 Klein 9B's weights are released under the FLUX Non-Commercial License . If you plan to sell anything you make, check each model's license first. **How to run them:** Most people use **ComfyUI**, which is the most flexible option and gets support for new models first. **Forge** and **Automatic1111** are simpler for Stable Diffusion models. **Quick answer:** If you have a strong GPU, start with FLUX. If you need commercial rights or good text in your images, use Qwen-Image. If you want the most add-ons, use Stable Diffusion.

Sources were recorded as search results.

Claude Code (local) 5 recorded answers
Flux · Stable Diffusion
Read answer 1 · 18:41 UTC

There isn't one clear winner, but these are the open-weight models people most often recommend right now: **Top picks** - **FLUX.2 [dev] (Black Forest Labs)**: The safest all-round choice. It gives high quality, consistent results at high resolution and has a large community of LoRAs and tools. Check the license: the "dev" weights are non-commercial unless you buy a commercial license. - **Qwen-Image (Alibaba)**: Possibly the best open option today, especially for **text inside images** and editing. Version 2.0 combines image generation and editing in one model and outputs at native 2K. It uses the permissive Apache 2.0 license. One source says a newer 2.1 release now tops the open-weight leaderboards. I couldn't confirm that, so check Hugging Face. - **HunyuanImage 3.0 (Tencent)**: Strong with long, complicated prompts. It's very large, so it needs serious GPU hardware. **Other good options** - **Stable Diffusion XL / 3.5 (Stability AI)**: Older and lower quality than the models above, but it has the biggest collection of fine-tunes, LoRAs and ControlNets, and it runs on modest GPUs. - **Sana (NVIDIA)**: Lightweight. It makes sharp images on an 8GB graphics card. **How to run them:** Use **ComfyUI** for the most control, or **Forge** or **InvokeAI** for something easier. **My recommendation:** Start with Qwen-Image if you want a permissive license or legible text in images. Use FLUX.2 [dev] for the best overall quality and community support. Choose SDXL if your GPU has 8–12GB of memory. Rankings in this area change every few months, and many "best of 2026" articles are vendor marketing. Some even list closed models like FLUX1.1 Pro as open source. Check the [Artificial Analysis](https://artificialanalysis.ai) or LMArena text-to-image leaderboards before you decide. Sources: - [Thunder Compute – best open source image generation models](https://thundercompute.com/blog/best-open-source-image-generation-models) - [Floyo – best open source image generation models in 2026](https://www.floyo.ai/blog/best-open-source-image-generation-models-in-2026) - [DEV – best open source AI image generators to self-host in 2026](https://dev.to/moksh/best-open-source-ai-image-generators-to-self-host-in-2026-36i4) - [Pixazo – top open source image generation models](https://pixazo.ai/blog/top-open-source-image-generation-models) - [WaveSpeed – best open source image models 2026](https://wavespeed.ai/landing/models/best-open-source-image-models-2026) - [SiliconFlow – best open source image generation models](https://www.siliconflow.com/articles/en/best-open-source-image-generation-models-2025) - [BentoML – guide to open source image generation models](https://origin.bentoml.com/blog/a-guide-to-open-source-image-generation-models)

How sources were used was not recorded.

    Flux · Qwen-Image · HunyuanImage · Z-Image · Sana
    Read answer 2 · 18:41 UTC

    There isn't one single best open-source image generator. These are the strongest choices right now: **Top picks** - **FLUX.2 (Black Forest Labs)** is the most commonly recommended. It produces high-quality, consistent images at high resolution, and handles text and hands well. FLUX.2 Dev is a good default if you're running it yourself. Note that the "Dev" weights carry a non-commercial license, so check the terms if you plan to use it for business. - **Qwen-Image (Alibaba)** is very strong at rendering text inside images and is under the permissive Apache 2.0 license. One guide reports that a newer 2.1 release went to the top of the open-weight leaderboard in September 2026. I couldn't confirm that, but this family is worth a close look, especially for commercial use. - **HunyuanImage 3.0 (Tencent)** handles complex, detailed prompts well. It's a very large model, so you'll need serious GPU hardware. **If your hardware is limited** - **Z-Image-Turbo** or **Sana** are good fits. Sana reportedly runs on an 8GB GPU. - **Stable Diffusion 3.5 / SDXL** are older, but they still have the largest collection of add-ons (LoRAs, ControlNets) and community support. **How to run them:** **ComfyUI** is the standard free interface for all of these. **Forge** and **InvokeAI** are easier to learn. **My recommendation:** Start with FLUX.2 Dev for overall quality, or Qwen-Image if you need a commercial-friendly license or accurate text in your images. Some lists also call Ideogram or FLUX Pro "open source," but those are closed, hosted models. Sources: - [DEV: Top 5 Open Source AI Image Models 2026](https://dev.to/techlatestnet/top-5-best-open-source-ai-image-generation-models-in-2026-tested-4edb) - [DEV: Best Open Source AI Image Generators to Self-Host 2026](https://dev.to/moksh/best-open-source-ai-image-generators-to-self-host-in-2026-36i4) - [Okara: Best Open Source Image Generation Models](https://okara.ai/blog/best-open-source-image-generation-models) - [Thunder Compute: Best Open Source Image Generation Models](https://thundercompute.com/blog/best-open-source-image-generation-models) - [Pinggy: Best Free Open-Source AI Image Generators to Self-Host](https://pinggy.io/blog/best_free_open_source_ai_image_generators_to_self_host/) - [SecondTalent: Top Open Source AI Image Generators](https://www.secondtalent.com/resources/top-open-source-ai-image-generators/) - [Pixazo: Top Open Source Image Generation Models](https://pixazo.ai/blog/top-open-source-image-generation-models)

    How sources were used was not recorded.

      Flux · Qwen-Image · Z-Image · Stable Diffusion · HunyuanImage
      Read answer 3 · 18:41 UTC

      There isn't one clear winner. It depends on your hardware and whether you need commercial rights. As of my last update (mid-2026), these are the ones I'd pick: **Best overall quality: FLUX (Black Forest Labs)** - FLUX.2 [dev] and FLUX.1 [dev] give top-tier realism, follow prompts well and render text well. - The catch: the [dev] weights are free to download but use a **non-commercial license**. FLUX.1 [schnell] is Apache 2.0, so you can use it commercially. **Best fully open option: Qwen-Image (Alibaba)** - Apache 2.0 license, so commercial use is fine. - It's excellent at text in images, including multiple languages, and it has a strong editing variant (Qwen-Image-Edit). - It's a large model, so you'll want a GPU with plenty of memory (VRAM) or a quantized (compressed) version. **Best for modest hardware: Z-Image Turbo (Alibaba Tongyi)** - Apache 2.0, fast, and runs well on consumer GPUs with about 12–16 GB of VRAM. Image quality is very good for its size. **Biggest ecosystem: Stable Diffusion XL / SD 3.5 (Stability AI)** - It's older and not the best on raw quality, but it has the most community fine-tunes, LoRAs (small add-on style or character models), ControlNets (tools for guiding pose or layout) and tutorials. Sites like Civitai host thousands of variants. **Also worth a look:** HunyuanImage (Tencent) and HiDream-I1 (MIT license). **How to run them:** the models need a front-end app. - **ComfyUI** is the most powerful and supports new models first. - **Forge** or **InvokeAI** are easier to use. - **Fooocus** is the simplest way to start with SDXL. **Quick pick:** - Best images, personal use: **FLUX** - Commercial use or text in images: **Qwen-Image** - An 8–12 GB GPU: **Z-Image Turbo** or **SDXL** New models come out every few months. Before you commit, check the current rankings on the Artificial Analysis or LMArena image leaderboards.

      How sources were used was not recorded.

        Flux · Qwen-Image · Z-Image · HunyuanImage · Stable Diffusion · Sana
        Read answer 4 · 18:41 UTC

        There's no single winner. Which model is best depends on your hardware, what you're making, and whether you need commercial-use rights. Here are my picks as of late 2026: **Best overall: FLUX.2 [dev]** (Black Forest Labs) - It gives the most consistent results, especially for photorealism and high-resolution images, and it has strong community support. - The license allows non-commercial use only. To sell your images you need a paid license, or you can use FLUX.1 [schnell], which is fully open under Apache 2.0. **Best truly open license: Qwen-Image** (Alibaba) - It's under Apache 2.0, so commercial use is fine. It's very good at putting readable text in images and at editing images. - Reports say version 2.1, released in September 2026, went straight to the top of the open-weight leaderboards. I haven't confirmed that myself. - Alibaba's **Z-Image Turbo** is a fast, lighter alternative. **Best for long, complex prompts: HunyuanImage 3.0** (Tencent) - It's very large and needs serious GPU power. **Best ecosystem: Stable Diffusion XL / 3.5** (Stability AI) - It's no longer the top model on quality. But it has by far the most add-ons, like LoRAs, ControlNets and community fine-tunes, and it runs on modest GPUs. **Best for weak hardware: Sana** (NVIDIA) - It produces sharp images on an 8GB graphics card. **Software to run them:** **ComfyUI** is the most powerful and supports every new model quickly. **Forge** or **InvokeAI** are easier to learn. **Fooocus** is the simplest way to start with SDXL. **My recommendation:** If you have a 16GB+ NVIDIA GPU, install ComfyUI and start with FLUX.2 [dev] for personal use, or Qwen-Image if you plan to sell what you make. If your GPU is smaller, try SDXL or Sana. One caution: many "open source" models are really "open weights," meaning you can download them but the license restricts how you use them. Read the license before using one commercially. For the latest rankings, check the open-weight filter on the Artificial Analysis Text-to-Image Arena. This field changes every few months. Sources: - [Pinggy – Best Free & Open-Source AI Image Generators to Self-Host](https://pinggy.io/blog/best_free_open_source_ai_image_generators_to_self_host/) - [Floyo – Best open source image generation models in 2026](https://www.floyo.ai/blog/best-open-source-image-generation-models-in-2026) - [DEV – Best open source AI image generators to self-host in 2026](https://dev.to/moksh/best-open-source-ai-image-generators-to-self-host-in-2026-36i4) - [DEV – Top 5 open-source AI image generation models in 2026](https://dev.to/techlatestnet/top-5-best-open-source-ai-image-generation-models-in-2026-tested-4edb) - [BentoML – A guide to open source image generation models](https://origin.bentoml.com/blog/a-guide-to-open-source-image-generation-models) - [Okara – Best Open Source Image Generation Models in 2026](https://okara.ai/blog/best-open-source-image-generation-models) - [WaveSpeed – Best open source image models 2026](https://wavespeed.ai/landing/models/best-open-source-image-models-2026)

        How sources were used was not recorded.

          Flux · Stable Diffusion
          Read answer 5 · 18:41 UTC

          For most people the best pick right now is **FLUX.2 [dev] from Black Forest Labs**. It's the open model reviewers recommend most often. Images are sharp and high-resolution, text comes out right, and it handles hands and anatomy well. **Strong alternatives:** - **Qwen-Image (Alibaba):** Best if you also want to edit images. It generates and edits in one model at native 2K resolution, and its license (Apache 2.0) is more permissive than FLUX's non-commercial dev license. One recent source says version 2.1 is now at the top of the open-weight rankings, but I haven't confirmed that. - **HunyuanImage 3.0 (Tencent):** Best for long, complicated prompts. It's very large, so it needs a lot of GPU memory. - **Stable Diffusion XL / SD 3.5 (Stability AI):** The biggest ecosystem of add-ons like LoRAs, ControlNets and community fine-tunes. Good if you want to customize a lot. - **Sana (NVIDIA):** Best if your graphics card is weak. It makes good images with about 8GB of VRAM. **Practical tips:** - Run these with **ComfyUI**, which is the most flexible, or with **Forge** or **InvokeAI** if you want something easier. - Check the license if you'll use the images commercially. "Open weights" doesn't always mean free for commercial use. FLUX dev and Ideogram's open model have restrictions, while Qwen-Image doesn't. - If you have about 12GB of VRAM or less, use quantized versions (GGUF/FP8) of FLUX or Qwen. **Quick recommendation:** For the best quality, use FLUX.2 [dev]. For commercial use or editing, use Qwen-Image. On a low-end GPU, use Sana or SDXL. Sources: - [Floyo – Best open source image generation models in 2026](https://www.floyo.ai/blog/best-open-source-image-generation-models-in-2026) - [Pinggy – Best Free & Open-Source AI Image Generators to Self-Host](https://pinggy.io/blog/best_free_open_source_ai_image_generators_to_self_host/) - [Thunder Compute – Best open source image generation models](https://thundercompute.com/blog/best-open-source-image-generation-models) - [DEV – Best open source AI image generators to self-host in 2026](https://dev.to/moksh/best-open-source-ai-image-generators-to-self-host-in-2026-36i4) - [WaveSpeed – Best open source image models 2026](https://wavespeed.ai/landing/models/best-open-source-image-models-2026) - [BentoML – A guide to open source image generation models](https://origin.bentoml.com/blog/a-guide-to-open-source-image-generation-models) - [SiliconFlow – Best Open Source Image Generation Models 2026](https://www.siliconflow.com/articles/en/best-open-source-image-generation-models-2025)

          How sources were used was not recorded.

            Codex (local) 5 recorded answers
            Flux
            Read answer 1 · 18:41 UTC

            **My pick for most people is Alibaba’s [Z-Image-Turbo](https://huggingface.co/Tongyi-MAI/Z-Image-Turbo).** It offers a strong balance of realistic images, speed, and manageable hardware requirements. It uses the permissive Apache 2.0 license, and its developers say it fits on a GPU with 16 GB of VRAM. Two alternatives worth considering: - **[Black Forest Labs’ FLUX.2 klein 4B](https://github.com/black-forest-labs/flux2)** if you want both image generation and editing. This particular version is Apache 2.0 licensed; other FLUX versions have different terms. - **[Alibaba’s Qwen-Image-2512](https://huggingface.co/Qwen/Qwen-Image-2512)** for detailed images and text-heavy designs, also under Apache 2.0. What computer or graphics card will you run it on? That can change the best choice considerably.

            How sources were used was not recorded.

              Flux
              Read answer 2 · 18:42 UTC

              **My pick is Black Forest Labs’ FLUX.2 [klein] 4B** for the best practical balance of image quality, speed, and running locally. It supports generation and editing, uses the permissive Apache 2.0 license, and can run with around 13 GB of GPU memory. [Model details](https://huggingface.co/black-forest-labs/FLUX.2-klein-4B) To use it, I’d choose **ComfyUI**, a free, open-source application with extensive control over image workflows. Its main drawback is a learning curve: the interface uses connected nodes. [ComfyUI](https://github.com/Comfy-Org/ComfyUI) Be specific about the model version when downloading—other FLUX models have different licenses. What computer or GPU will you be using? That could change my recommendation.

              How sources were used was not recorded.

                Flux
                Read answer 3 · 18:42 UTC

                **My pick is Alibaba’s Qwen-Image-2512 if image quality is your priority.** It’s especially strong for realistic people, fine detail, and readable text within images, and it uses the permissive Apache 2.0 license. [Model and examples](https://huggingface.co/Qwen/Qwen-Image-2512) Two alternatives are worth considering: - **Z-Image-Turbo from Alibaba:** My choice for fast, everyday image generation, with strong photorealistic results. [Details](https://github.com/Tongyi-MAI/Z-Image) - **FLUX.2 [klein] 4B from Black Forest Labs:** A practical choice for local generation **and editing** on consumer hardware. This specific version is Apache 2.0; the larger 9B and dev versions have noncommercial licenses. [Details](https://github.com/black-forest-labs/flux2) If you’re running it locally, your graphics card and available memory will heavily influence which is best for you.

                How sources were used was not recorded.

                  Flux
                  Read answer 4 · 18:42 UTC

                  **My pick is Alibaba’s Qwen-Image-2512** if image quality is your priority. It’s a strong option for realistic people, detailed scenes, and readable text inside images, with downloadable weights under the permissive Apache 2.0 license. [Model and examples](https://huggingface.co/Qwen/Qwen-Image-2512). **For a more practical starting point on your own computer, consider Black Forest Labs’ FLUX.2 [klein] 4B.** Its smaller model supports both generation and editing, and that specific version is Apache 2.0 licensed. Other FLUX versions have different restrictions, so check the exact variant. [Models and licenses](https://github.com/black-forest-labs/flux2). What computer or graphics card would you run it on? That can change which is the best choice for you.

                  How sources were used was not recorded.

                    Flux
                    Read answer 5 · 18:42 UTC

                    **My pick is Alibaba’s Qwen-Image-2512** if image quality is your priority. It’s particularly strong at realistic people, fine detail, and readable text within images, and it uses the permissive Apache 2.0 license. [Model details](https://huggingface.co/Qwen/Qwen-Image-2512) For faster generation on your own computer, **Black Forest Labs’ FLUX.2 [klein] 4B** is the more practical choice. It supports generation and editing, with Apache 2.0 licensing specifically for the 4B version. [Model details](https://github.com/black-forest-labs/flux2) Your hardware matters: what computer or graphics card would you run it on?

                    How sources were used was not recorded.

                      Gemini 5 recorded answers
                      Flux · Stable Diffusion · Z-Image
                      Read answer 1 · 18:39 UTC

                      The consensus best open-source (open-weights) AI image generator model is **FLUX** by **Black Forest Labs**, while **Stable Diffusion** by **Stability AI** remains the top choice if you want the largest ecosystem for customization and low hardware requirements. The top options depend on your technical setup and needs: --- ### 1. The Best Overall Quality: **FLUX.1 & FLUX.2 (by Black Forest Labs)** Created by the original engineers who developed Stable Diffusion, the FLUX family is currently the gold standard for open-weight image generation. It handles complex prompts, photorealism, finger/hand details, and legible text within images far better than older open models. * **FLUX.1 Schnell:** Fully open source under an **Apache 2.0 license** (free for commercial use). Optimized for high-speed, 4-step generation. * **FLUX.1 Dev / FLUX.2 Dev:** Open-weight models free for non-commercial/personal use. Delivers commercial-grade image quality. * *Hardware needed:* Requires a higher-end GPU (ideally 12GB to 16GB+ VRAM). --- ### 2. The Best Ecosystem & Customization: **Stable Diffusion XL & SD 3.5 (by Stability AI)** While FLUX beats standard Stable Diffusion in out-of-the-box quality, Stable Diffusion (especially **SDXL** and **Stable Diffusion 3.5 Large**) remains an industry workhorse. * **Why choose it:** Unmatched community support. Platforms like Civitai host tens of thousands of free custom styles, fine-tunes, ControlNets, and LoRAs specifically trained for Stable Diffusion. * **Hardware friendly:** Runs comfortably on modest GPUs with 6GB–8GB of VRAM. --- ### 3. Best for Speed & Low VRAM: **NVIDIA Sana** or **Z-Image** If you don't have a high-end desktop GPU or want near-instant generations on a laptop: * **NVIDIA Sana:** A lightweight model capable of generating high-resolution (1024x1024) images in under a second on standard consumer hardware. * **Z-Image:** Highly efficient open model tailored for fast, low-latency generation. --- ### How to Actually Run Them These underlying models (weights) need a front-end software interface to run locally on your computer: * **InvokeAI:** The cleanest, most user-friendly interface for digital artists and beginners. * **ComfyUI:** The industry standard for power users—node-based, modular, and fast. * **AUTOMATIC1111 / Forge:** Popular web-based dashboards with massive plugin support.

                      How sources were used was not recorded.

                      Flux · Stable Diffusion
                      Read answer 2 · 18:39 UTC

                      The best open-source (open-weight) AI image generator currently available is **FLUX** by **Black Forest Labs**, with **Stable Diffusion** by **Stability AI** as the runner-up depending on your hardware and customization needs. --- ### 1. Overall Best Image Quality & Text: **FLUX (Black Forest Labs)** Created by the original engineers behind Stable Diffusion, the **FLUX.1** model family is widely considered the state of the art for open models. * **Why it wins:** Unmatched photorealism, superior handling of human hands/anatomy, and best-in-class text rendering (spelling words on signs/labels accurately). * **Variants:** * **FLUX.1 [dev]:** Open-weight for personal/non-commercial use; offers maximum image quality. * **FLUX.1 [schnell]:** Fully open-source under an Apache 2.0 license, optimized for fast 4-step generation. * **Catch:** Requires a relatively beefy GPU (12GB+ VRAM recommended for running locally without quantization). --- ### 2. Best for Ecosystem & Budget GPUs: **Stable Diffusion (Stability AI)** While FLUX leads in raw quality out of the box, **Stable Diffusion XL (SDXL)** and **Stable Diffusion 3.5** remain community favorites. * **Why choose it:** * **VRAM Efficiency:** Runs smoothly on mid-range GPUs (8GB VRAM or less using SDXL or SD 1.5). * **Ecosystem:** Tens of thousands of custom community fine-tunes, LoRAs (style modifiers), and ControlNets are available on sites like **Civitai**. * **Styles:** Excellent for anime, stylized concept art, and fine-grained control via custom workflows. --- ### Best Free Interfaces to Run Them Locally If you want to run these models on your own PC, you will need a software front-end: * **ComfyUI:** The industry standard for power users. Uses a node-based interface that gives you complete control over your generation pipeline. * **Forge (WebUI):** A user-friendly, browser-based interface tailored for straightforward text-to-image workflows. * **InvokeAI:** An open-source interface designed specifically for digital artists and creative professionals, featuring canvas-based editing tools. --- ### Recommendation Summary: * Choose **FLUX.1 [dev] / [schnell]** if you have a powerful NVIDIA GPU (12GB+ VRAM) and want the highest prompt accuracy and photorealism. * Choose **SDXL** if you have a mid-tier GPU or want to experiment with thousands of community-built visual styles and LoRAs.

                      How sources were used was not recorded.

                      Flux · Stable Diffusion · Ideogram · Qwen-Image · HunyuanImage
                      Read answer 3 · 18:39 UTC

                      The answer depends on whether you are looking for the underlying **AI model** (open-weights) or the **software interface** to run on your computer. --- ### Best Open-Weight AI Models 1. **FLUX series (by Black Forest Labs)** – *Best Overall Image Quality* * **Why it’s the best:** Developed by key creators of original Stable Diffusion, the **FLUX** model family (such as FLUX.1 `[dev]` and `[schnell]`) offers industry-leading photorealism, text rendering, and prompt adherence. It handles complex lighting, hands, and detailed composition better than most proprietary competitors. * **Licensing:** FLUX.1 `[schnell]` is open under Apache 2.0 (commercial friendly), while FLUX.1 `[dev]` is open for non-commercial/research use. 2. **Stable Diffusion 3.5 (by Stability AI)** – *Best for Customization & Ecosystem* * **Why it’s great:** Stable Diffusion 3.5 (Large and Medium variants) offers solid prompt adherence and rendering. Its biggest advantage is the massive community ecosystem: thousands of custom fine-tunes, LoRAs (style adaptations), and ControlNet plugins are readily available. * **Licensing:** Stability AI’s Community License allows free commercial use for small businesses/creators under certain revenue thresholds. 3. **Ideogram 4.0 & Qwen-Image** – *Honorable Mentions* * **Ideogram 4.0** (Open Weights): Exceptionally strong at generating clear, accurate text within images. * **Qwen-Image / HunyuanImage:** Excellent open alternatives for complex prompt adherence and multilingual text generation. --- ### Best Open-Source Interfaces to Run Them Locally To use these models on your own GPU, you'll need a user interface: * **Invoke (InvokeAI):** The best option for artists and designers. It offers a clean, Photoshop-like canvas interface, inpainting/outpainting tools, and local project management. * **ComfyUI:** The gold standard for advanced power-users. It uses a visual node-based workflow allowing extreme control over inference pipelines, generation speed, and batch processing. * **AUTOMATIC1111 (Stable Diffusion WebUI):** The classic, easy-to-install browser UI with extensive extensions and community tutorials. --- ### Recommendation * If you want the **absolute best visual output** with a modern GPU, download **FLUX.1 [dev]** or **FLUX.1 [schnell]**. * If you want a **free desktop app** to run models locally without coding, pair **Invoke** or **Pinokio** (one-click installer) with **FLUX** or **Stable Diffusion 3.5**.

                      How sources were used was not recorded.

                      Flux · Stable Diffusion
                      Read answer 4 · 18:39 UTC

                      The current top recommendation for an open-source (open-weights) AI image generator is **FLUX.1** by **Black Forest Labs**. ### 1. Overall Best Image Quality & Text: **FLUX.1 (by Black Forest Labs)** Created by the original team behind Stable Diffusion, FLUX.1 leads open-weights image generation in photorealism, prompt accuracy, and rendering legible text. * **FLUX.1 [schnell]:** Fully open-source under an **Apache 2.0 license** (free for personal and commercial use). It runs very fast (in just 4 steps) while producing great results. * **FLUX.1 [dev]:** Open-weights (non-commercial use). It takes slightly longer to generate images but delivers higher detail and fidelity. * *Hardware requirement:* Requires a dedicated GPU (ideally 12GB+ VRAM for full precision, though quantized 8-bit/4-bit versions run smoothly on 8GB VRAM). ### 2. Best for Custom Fine-Tuning & Lower VRAM: **Stable Diffusion XL (SDXL)** While Stability AI has released newer models like **Stable Diffusion 3.5**, **SDXL** remains the king of community customization. * **Why pick it:** It has the largest ecosystem of free community-trained fine-tunes, LoRAs (style adapters), and ControlNet tools available on sites like Civitai. * **Hardware requirement:** Runs well on budget or mid-range GPUs with as little as 6GB–8GB VRAM. --- ### Best Free Software to Run Them Locally To generate images on your own computer, pair these models with an open-source local interface: * **InvokeAI:** The most user-friendly, artist-focused interface. It feels like a real digital art application with canvas tools and layer support. * **ComfyUI:** The industry standard for power users. It uses a visual node-based workflow, allowing fine-grained control over speed, memory usage, and image manipulation. * **WebUI Forge (or Automatic1111):** Traditional web interfaces that offer a straightforward, tab-based layout with broad extension support.

                      How sources were used was not recorded.

                      Flux · Stable Diffusion · Qwen-Image
                      Read answer 5 · 18:39 UTC

                      When choosing an open-source or open-weights AI image generator, the best option depends on whether you prioritize image quality, commercial usage freedom, or customizability (like LoRAs and fine-tuning). The top options currently available: --- ### 1. **Best Overall Quality & Text Rendering: FLUX (Black Forest Labs)** Developed by former Stability AI researchers, the **FLUX series** (such as FLUX.1 and FLUX.2) is widely considered the state-of-the-art for open-weight image generation. * **Strengths:** Unrivaled prompt adherence, photorealism, and clear text/typography rendering inside generated images. * **Which variant to use:** * **FLUX [schnell]:** Fully open-source under an **Apache 2.0 license** (free for commercial use). Extremely fast. * **FLUX [dev]:** Higher quality open-weights version for non-commercial/research use (commercial licenses available). * **Hardware requirements:** High VRAM (ideally 16GB–24GB+ GPU for smooth local generation, though quantized versions run on lower-spec hardware). --- ### 2. **Best Ecosystem & Customization: Stable Diffusion (Stability AI)** While newer architectures exist, **Stable Diffusion 3.5** and **SDXL** remain the standard for customization. * **Strengths:** Huge open-source community support. If you want to train custom characters, use ControlNet for precise pose/composition matching, or download specialized artistic styles, the SD ecosystem is unmatched. * **Which variant to use:** * **SD 3.5 Large / Medium:** Modern Diffusion Transformer (DiT) base with improved anatomy and composition. * **SDXL 1.0:** Older, but has the largest library of community fine-tunes, LoRAs, and workflow plugins. --- ### 3. **Best Unrestricted Commercial Model: Qwen-Image (Alibaba)** * **Strengths:** High-performing open-weight models released under fully permissive open licenses (Apache 2.0). * **Best for:** Developers looking to integrate high-quality visual generation into commercial software without restrictive licensing fees. --- ### Recommended Software to Run Them To run these models locally on your own computer, free open-source UIs include: * **ComfyUI:** The node-based interface used by power users and pros for maximum workflow control. * **InvokeAI:** A polished, user-friendly canvas-style interface designed for artists and designers. * **SD WebUI Forge:** A streamlined version of the classic Automatic1111 interface optimized for faster generation and lower GPU VRAM usage.

                      How sources were used was not recorded.

                      Perplexity 5 recorded answers
                      Flux
                      Read answer 1 · 18:39 UTC

                      **Best overall: FLUX.2 [klein] 4B** from Black Forest Labs. It supports both image generation and editing, is designed to run on consumer GPUs, and its weights are Apache 2.0—so it’s a strong pick for local use and commercial projects. For a simple way to run it, try **ComfyUI**, an open-source app that supports FLUX and many other image models. Check your hardware and the model’s license before downloading: FLUX.2 [klein] 9B and FLUX.2 [dev], for example, have more restrictive terms.

                      How sources were used was not recorded.

                      Flux · Stable Diffusion
                      Read answer 2 · 18:39 UTC

                      **Best overall: FLUX 3 Image Klein**—it’s a strong all-around choice for image generation and editing, with open weights and versions designed to run on consumer hardware. Its FLUX Kommunity License allows free commercial use for individuals and companies under $5 million in annual revenue; check the terms if you’re above that threshold or need broader usage rights. If you specifically want a **permissive license for commercial use**, consider **FLUX.2 Klein 4B**, released under Apache 2.0. For a well-established alternative that runs on modest hardware, **Stable Diffusion 3.5** is also worth a look, though its Community License has revenue-based terms. “Open source” can mean different things here: these models generally provide **open weights**, but licenses and commercial rights vary.

                      How sources were used was not recorded.

                      Flux · Stable Diffusion
                      Read answer 3 · 18:39 UTC

                      **Best overall: FLUX.2** from Black Forest Labs—it’s a strong pick for image quality and prompt accuracy. For a fully permissive, easy-to-use option, choose **FLUX.2 Klein 4B**, whose weights are Apache 2.0 licensed. If you want the biggest library of community models, add-ons, and tutorials, **Stable Diffusion** remains a great alternative. “Open source” can mean different things here: many image models publish their weights but use licenses with commercial-use restrictions, so check the specific model’s terms before using it in a business.

                      How sources were used was not recorded.

                      Flux · Qwen-Image · Stable Diffusion
                      Read answer 4 · 18:40 UTC

                      **ComfyUI** is my best all-around recommendation for an open-source image-generation tool. It’s free, runs locally, and lets you choose among models rather than locking you into one. For the model, start with **FLUX.2** for strong image quality and prompt following. If commercial licensing matters, check the specific model variant’s terms: “open weights” doesn’t always mean unrestricted commercial use. **Qwen Image** is worth trying when readable text in images is important; **Stable Diffusion** remains a solid choice for its huge ecosystem of add-ons and community models.

                      How sources were used was not recorded.

                      Qwen-Image · Z-Image
                      Read answer 5 · 18:40 UTC

                      **Best overall: Qwen-Image-2.1**, especially if you want a strong open-weight model for both generating and editing images. It’s a good all-around pick; you can run it through **ComfyUI**, an open-source interface. If you mean *permissively licensed* open source for commercial use, look at **Z-Image Turbo** (Apache 2.0) instead. “Open source” can mean different things for AI models, so check the model’s license and hardware requirements before choosing.

                      How sources were used was not recorded.

                      q14

                      Which AI image generator is safest for licensing and copyright? #

                      Named first most often: Adobe Firefly (30 of 30 answers with these filters).

                      30 recorded attempts with these filters.

                      Read full answers to this question

                      ChatGPT 5 recorded answers
                      Adobe Firefly
                      Adobe Firefly
                      Adobe Firefly
                      Adobe Firefly
                      Adobe Firefly
                      Claude 5 recorded answers
                      Adobe Firefly
                      Adobe Firefly · Midjourney · Stable Diffusion
                      Adobe Firefly · Ideogram · Midjourney · Stable Diffusion
                      Adobe Firefly · Midjourney · Stable Diffusion
                      Adobe Firefly · Midjourney · Stable Diffusion
                      Claude Code (local) 5 recorded answers
                      Adobe Firefly · Canva · Midjourney · Stable Diffusion
                      Adobe Firefly · Midjourney · Stable Diffusion
                      Adobe Firefly · Midjourney · Stable Diffusion
                      Adobe Firefly · Midjourney · Stable Diffusion
                      Adobe Firefly · Imagen · Midjourney · Stable Diffusion
                      Codex (local) 5 recorded answers
                      Adobe Firefly
                      Adobe Firefly
                      Adobe Firefly
                      Adobe Firefly
                      Adobe Firefly
                      Gemini 5 recorded answers
                      Adobe Firefly
                      Adobe Firefly · Midjourney
                      Adobe Firefly
                      Adobe Firefly · Midjourney · DALL-E · Imagen
                      Adobe Firefly · Midjourney · Stable Diffusion
                      Perplexity 5 recorded answers
                      Adobe Firefly
                      Adobe Firefly
                      Adobe Firefly
                      Adobe Firefly
                      Adobe Firefly

                      q15

                      What AI image generator should a small business use? #

                      Named first most often: Canva (26 of 30 answers with these filters).

                      30 recorded attempts with these filters.

                      Read full answers to this question

                      ChatGPT 5 recorded answers
                      Canva · DALL-E · Adobe Firefly · Midjourney
                      Canva · Adobe Firefly
                      Canva · Adobe Firefly · Midjourney
                      Canva · Adobe Firefly · Ideogram
                      Canva · DALL-E · Adobe Firefly · Midjourney
                      Claude 5 recorded answers
                      Canva · Adobe Firefly · Imagen · Midjourney
                      Canva · Adobe Firefly · Ideogram · Imagen
                      Canva · Adobe Firefly · Ideogram · Imagen · Midjourney
                      Canva · Adobe Firefly · Midjourney · Imagen · Ideogram
                      Canva · Adobe Firefly · Midjourney · Imagen
                      Claude Code (local) 5 recorded answers
                      Adobe Firefly · Canva · Midjourney · Ideogram
                      Adobe Firefly · Canva · Midjourney · Imagen · Ideogram
                      Adobe Firefly · Canva · Midjourney · Imagen · Ideogram
                      Canva · Adobe Firefly · Midjourney · Imagen · Ideogram
                      Adobe Firefly · Canva · Midjourney · Imagen · Ideogram
                      Codex (local) 5 recorded answers
                      Canva · Adobe Firefly
                      Canva · Adobe Firefly
                      Canva · Adobe Firefly
                      Canva · Adobe Firefly
                      Canva · DALL-E · Adobe Firefly
                      Gemini 5 recorded answers
                      Canva · Adobe Firefly · Midjourney · Ideogram
                      Canva · Adobe Firefly · Ideogram · Midjourney
                      Canva · Adobe Firefly · Ideogram · Midjourney
                      Canva · Adobe Firefly · Ideogram · Midjourney
                      Canva · Adobe Firefly · Midjourney · Ideogram
                      Perplexity 5 recorded answers
                      Canva · Adobe Firefly
                      Canva · Adobe Firefly
                      Canva · Adobe Firefly
                      Canva · Adobe Firefly
                      Canva · Adobe Firefly · Midjourney