Viral prompt recipe makes GPT image models shoot 'unpublished' photojournalism
techhalla · x · 2026-10-06
Creator techhalla shares his 'GPT 2.5 Flare' prompt for GPT image models, directed to produce stills that read as unpublished photographs by a working photographer — optical capture, available physics, lived-in detail.
The example scene is hyper-specific: three 40-50-something Arizona trailer park neighbors on a faded 1980s floral couch lifted four meters by 50 white weather balloons, down to sunburns, goatees and dented beer cans. The combo of overly specific absurd-realistic scenes plus documentary photo language is the core trick; a matching Seedance prompt is included in the reply.
More from Multimodal
- Claude Opus 5.5 'codes video' with a Seedance API key to produce a Bosch-inspired film trailer — ZeroStateReflex · 2026-10-07
- Image-to-video model Prism with joint video-audio generation trends on Hugging Face, MIT licensed — FrancisRing · 2026-10-07
- iADD corrects DDPO theory: latter-timestep-only updates can harm diversity — Ashok Prasad Neupane · 2026-10-07
- Foresight lets streaming VLMs plan future perception without retraining, beating baselines by 9.5% — Ashok Prasad Neupane · 2026-10-07
- Qwen-Image-2.1 Pro lands on Runware: native 2K generation at $0.075 per image — aziz4ai · 2026-10-07
- Nano Banana 2.1 lands with better text rendering and multi-frame layout consistency — ColbyHawker · 2026-10-07