Image generation · Detailed review
Stability AI
Stable Diffusion 3.5 Large
« The open-source standard. Deploy at home for sovereignty or R&D experimentation. »
RadarOnAI score
/ 10 · in-house test
§ Walkthrough video · Stable Diffusion 3.5 Large
YouTube · officiel
§ Product preview
Score
7.8/10
weighted average
Price
Open weights
indicative
Specialty
Self-hosted
editorial verdict
Vendor
Stability AI
Self-hosted · ComfyUI · LoRA
Category
Image generation
8/9 on the ranking
§3.2+ Strengths
- 01Open weights (Apache 2.0)
- 02ComfyUI, ControlNet, LoRA ecosystem
- 03Self-hostable, data stays at your place
§3.3− Weaknesses
- 01Non-trivial setup (GPU required)
- 02Raw quality below Midjourney
§04How it positions
Stable against the segment average.
Tool score
7.8
/ 10
Segment average
8.7
average across 9 tools
Position
8/9
RadarOnAI ranking
§ Try Stable live
We use it in editorial. If image generation is part of your daily work and Stable Diffusion 3.5 Large ticks your use cases, an hour with it beats every review on the planet.
External link ↗ — opens the official site in a new tab.
§05Also worth comparing
Three other image generators to look at.
Google DeepMind · #1
Nano Banana 2 (Gemini 3.1 Flash Image)
9.4
/10
« Flash speed plus Gemini 3 intelligence. Plugs into Google's knowledge base for factual rendering. »
Speed + knowledgeRead the review →
Midjourney · #2
Midjourney V8 Alpha
9.3
/10
« The most beautiful, period. First pick for art direction and moodboards. »
AestheticsRead the review →
Google DeepMind · #3
Nano Banana Pro (Gemini 3 Pro Image)
9.2
/10
« The high-end mode kept for specialised tasks. Regenerate via the three-dot menu in Gemini. »
Pro qualityRead the review →