I have never gotten prompt mixing to work well, and I clearly need to practice more, but I believe mastering prompt mixing is how you can take AI art from kinda cool to extraordinary. This is the best I've managed with prompt mixing, and if anyone has advice on how I can improve, I would love to hear it.

Generated with NovelAI Diffusion Furry (Beta v1.3)

landscape normal res

prompt 1: {{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{fluttershy, pony, pegasus, cute}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}, Masterpiece, Ultra detailed, great composition, {{eerie and aesthetic}}, {dramatic lighting}, Mythological, high quality, intricate details, {{{intricate hair details}}}}, {{{{{beautiful detailed eyes}}}}}, {{{{{beautiful detailed background}}}}}

prompt 2: {Full Body}, depth of field, detailed eyes, fine detailed hair, dynamic shadows, dynamic angle, {{Extremely detailed eyes}}

prompt 3: {{{deep eyes}}}, {{{gemstone eyes}}}, {{{{{{{character focus}}}}}}}, {light diffusion}, {{{{harmonious}}}}, {best illustration}, {dramatic shadow}, {lightfast}, 1girl, {{{{iridescent}}}}

prompt 4: {1girl}, {solo}, {{{dynamic angle}}}, {{dynamic pose}}, {{{masterpiece}}}, illustration, {{{{{beautiful detailed sky}}}}}, {{{{{beautiful-detailed eyes}}}}}, {{cinematic light}}, ultra detailed, {{{{{disheveled hair}}}}}, jewelry teal eyes, {{{frills}}}, standing, day, floating cloud, {{{beautiful detailed cloud}}}, {glow white particles}, from side, serious, parted lips

undesired:
NONE
negatives: simple background, lowres, bad anatomy, bad hands, disfigured face, cropped, worst quality, low quality, normal quality, jpeg artifacts, extra leg, ugly, watermark, missing limbs, extra limbs, mutilated, mutation

yes add quality tags

steps: 28
scale: 7
seed: 4126406664

sampler:
do not use SMEA
Euler (not ancestral)