Hands-on with ChatGPT Sketch and Images 2.5: rough doodles become posters
APPSO · wechat · 2026-09-09
APPSO tested ChatGPT's new Sketch canvas and Images 2.5. Typing @Sketch opens a canvas where users draw outlines and add text prompts; a crude movie-poster sketch was understood by the model, which offered 5 artistic directions and generated psychological-thriller and retro-poster variants, while a rough still-life sketch became a proper photo with wood grain and side lighting.
Other tests include generating 8 camera angles of the Titanic bow scene and a 16-frame vlog storyboard to check character/clothing consistency across shots, plus long-prompt experiments with flat vector illustration, crayon-doodle overlays on photos, and early Source-engine-style game screenshots, where the model followed constraints on geometry, color palette and negative space well. Full prompts are included.
The takeaway: sketches can be rough while requirements stay specific—users can decide composition earlier and actively shape the image rather than passively receiving results. Note that local edits aren't always precise, so details and text should be checked before delivery.
Related event: ChatGPT quietly upgrades to GPT-Image-2.5 with new Sketch feature(20 posts)→
More from Multimodal
- Creator builds Snorricam video series with Seedance 2.5, sharing prompts — LudovicCreator · 2026-10-27
- Gradium launches free webpage TTS Chrome extension powered by its 40M-parameter Phonon model — mattturck · 2026-10-27
- Vortex of Infinite Recursion prompt goes viral: images that spiral endlessly into themselves — azed_ai · 2026-10-27
- 170+ AI Creators Made One Collective Film, Premiering November 6 — Loo_Atreides · 2026-10-27
- ComfyUI on 12GB VRAM: measured peaks show offloading turns 24s into 11 minutes — Impossible-Mail-4619 · 2026-10-27
- Midjourney creator showcases --sref style reference outputs: gothic character art — OVolosin82152 · 2026-10-27