I have been using DALL·E 3 within ChatGPT Plus for almost two years, and while GPT has improved significantly (now GPT-5), the DALL·E image generation model has not received any major updates since its release in October 2023.
As a paying subscriber, I expect the included image generation tool to evolve at the same pace as the text models. DALL·E 3 still struggles with:
– Preserving exact character designs from reference images.
– Maintaining consistent art styles (e.g., matching a specific splash art style).
– Accurately reproducing colors and fine details without distortion.
Competing services (MidJourney, Leonardo AI, Stable Diffusion) have made frequent and substantial improvements in this period, and DALL·E is falling behind.
I kindly request that OpenAI prioritize a major update to DALL·E — whether it’s an improved DALL·E 3 or the release of DALL·E 4 — with better reference fidelity, style matching, and color accuracy. This would ensure the image generation tool matches the quality and progress of GPT-5.
Thank you for your attention to this matter.