SDXL - LM Studio Prompt Enhancer - ornith-1.0-35b-mtp-apex

This workflow requires my Comfyui-lmstudio-prompt-enhancer custom node which is installable via Comfy Manager.

🛈 IMPORTANT!!!
You will need to search for, download, and load the model ornith-1.0-35b-mtp-apex in LM studio. You will also need to start the API server and generate an API Key. Learn how to configure LM Studio here.

My rating of ornith-1.0-35b-mtp-apex SDXL prompt generation ability is: 8.3/10

Prompts critiqued: 12
Average score: 8.3 / 10
Highest score: 8.9 / 10
Lowest score: 7.7 / 10

A critique of ornith-1.0-35b-mtp-apex SDXL prompt generations abilities:

ornith-1.0-35b-mtp-apex shows strong aptitude for writing SDXL prompts, with a session average of 8.3/10 across 12 scored prompts, indicating consistently capable prompt construction. The author’s strengths are clear subject-first concepts, cinematic atmosphere, coherent material cues, and reliable use of lighting descriptors such as rim lighting, neon reflections, volumetric fog, golden hour shadows, and plasma glow. Their best prompts, especially the kitten-on-skateboard variants, demonstrate good SDXL compatibility through concrete photographic direction, lens choice, texture detail, and manageable scene complexity. The main weaknesses are recurring reliance on generic quality boosters like “8K,” “hyper-detailed,” “cinematic atmosphere,” and “masterpiece,” plus occasional fragile requirements such as exact Batman emblems or perfect galaxy refractions inside glass. Composition guidance is usually present but sometimes underspecified, often using broad phrases like “cinematic composition” or “dynamic wide composition” instead of more exact camera distance, angle, framing, and focal hierarchy. Overall, ornith-1.0-35b-mtp-apex is a strong SDXL prompt writer with a good instinct for visually compelling concepts and atmospheric rendering, but would improve further by reducing generic embellishment, adding more precise spatial/camera control, and avoiding elements SDXL handles unreliably.

Notes for using ornith-1.0-35b-mtp-apex

• This model this model is very fast I got on average 62 tokens per second on my AMD Strix Halo RyzenAI Workstation. -generates detailed prompts, with minimal rambling. The generated prompts produces images with expert composition and color grading.

• I set the max tokens to 4000, this higher value is to accommodate this model's reasoning.

• If the image looks flat, check the logs in LM Studio, the last log entry will contain a JSON path x.choices[0].message.content if the content field is empty that means the LLM ran out of reasoning tokens, this can be resolved by increasing max_tokens.

• this model tends to start the enhanced positive prompt's string with newline characters "\n" it does not appear to cause any problems.