Breaking

OpenAI's GPT-Image 2.5 Wins Praise for Polished Game and App Interface Renders in Wider Rollout

September 8, 2026 at 21:26 EDT

Foundation Models
Media Generation

OpenAI has started rolling out ChatGPT Images 2.5, an upgraded image-generation model that developers say produces unusually clean, text-accurate interface art — including game menus and heads-up displays — while cutting rendering latency by as much as 50 percent. The update reached ChatGPT, ChatGPT Work, and Codex users across desktop, mobile, and web on September 8, 2026, and OpenAI paired it with two corresponding models in the Images API, GPT-Image-2.5 Flare and GPT-Image-2.5 Sunburst, according to OpenAI's announcement.

Sept 8, 2026 · OpenAI

GPT-Image 2.5 Sharpens AI-Generated Interfaces

OpenAI's upgraded image model cuts generation latency by up to 50% and wins early praise for rendering polished, text-accurate game and app interfaces — at unchanged API prices.

-50%

Latency vs. prior model

$30/1M

Output tokens, price unchanged

xhigh / max

New quality tiers

ChatGPT + API + Codex

Rollout scope

Generation time, same request

GPT-Image-2

baseline

GPT-Image-2.5 Flare

-50%

Flare renders in roughly half the time of its predecessor, at the same per-token price.

Well received

Praised for clean, legible game and app interface mockups
Independent tests clock it 2-4x faster than GPT-Image-2
Quality and speed gains at no added per-token cost

Early friction

Per-request token consumption not fully documented
Third-party pricing calculators lag the new models
No batch-generation support at initial launch

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.