First-shot blind test of P-H3, Flux3-Klein, Moss and Krea2-Raw on four hard prompts
AxonkaiLab · reddit · 2026-11-05
The author ran a zero-cherry-picking, first-shot comparison of four image models — P-H3, Flux3-Klein, Moss and Krea2-Raw — on a custom ComfyUI backend with no upscalers or inpainting. Four demanding prompts tested spatial reasoning, counting, lighting constraints and anatomy: a rainy bus stop portrait, a studio wristwatch still life, a coastal courtyard at dusk, and a fox crossing a woodland stream. All first-shot outputs are published slide by slide for comparison.
More from Multimodal
- SenseNova 6.8 Flash launches with long-term memory, animated PPTs and editable Blender projects — alifcoder · 2026-11-05
- Prompt share: cinematic Western RPG character renders with Unreal Engine 5 look — azed_ai · 2026-11-05
- AI-generated Madonna-style photos show off cinematic lighting realism — egeberkina · 2026-11-05
- Lip-syncing a 3:20 AI music video on one RTX 5090: LTX + Wan 2.2 workflow — keto_brain · 2026-11-05
- Korean AI short drama "Love at First Shot" trailer made with Seedance — Beginning_Victory597 · 2026-11-05
- 744 words and under $10: AI-generated film clip sparks filmmaking debate — umesh_ai · 2026-11-05