ChatGPT Images 2.5
Points and comments are a snapshot, not live.
OpenAI launches ChatGPT Images 2.5 with sharper details, faster generation, and a Sketch feature.
OpenAI released ChatGPT Images 2.5, a state-of-the-art image model offering sharper details, more precise editing, 50% lower latency than Images 2.0, and better multi-turn consistency. New product features include Sketch for drawing directly in ChatGPT, templates for popular formats like flyers, and comment-based editing. In the API, two models are available: GPT-Image-2.5 Flare for general use and GPT-Image-2.5 Sunburst for detailed creative work. The model improves reference photo fidelity, natural lighting, and texture. Pricing and availability cover all ChatGPT tiers, ChatGPT Work, and Codex users on desktop, mobile, and web.
Safety measures include prompt and image checks, C2PA metadata, and invisible watermarking. The model is rolling out today.
What commenters are saying
Commenters are split: some appreciate the technical improvements, while many criticize the model's persistent flaws, like distorted fingers, unnatural shadows, and loss of fine detail in composite images (e.g., the party photo).
A recurring theme is the association of ChatGPT images with "AI menu slop" seen in real businesses, which some find off-putting. Others argue small shops benefit from accessible design tools, but many prefer human-made signs. A few note the irony that OpenAI's own promotional examples contain obvious errors, suggesting the team may lack attention to detail. Overall, sentiment ranges from cautious optimism to disappointment that basic issues remain unsolved.