GPT Image 2.5 Sunburst

A press release dated September 11-12, 2026 describes GPT Image 2.5 as OpenAI's latest image generation platform, splitting image creation into two purpose-built models: Flare for speed-first workflows and Sunburst for precision-first creative production. Sunburst is positioned for maximum editing control and productio The release frames Sunburst as the variant of choice for precision editing rather than throughput, complementing Flare's speed-first design. It lists Sunburst alongside a Sketch drawing tool and multi-turn comment-based editing as part of the GPT Image 2.5 launch bundle. The source is a syndicated press-release wire it

Tech Insider reports that on September 8, 2026, OpenAI released GPT Image 2.5 Sunburst and GPT Image 2.5 Flare for image generation and editing through the Image API and the Responses API image generation tool, quoting the company's developer changelog directly. The article distinguishes this developer-facing pair of models from the consumer ChatGPT Images 2.5 release, which shipped alongside a new Sketch drawing mode. It frames the launch as one of OpenAI's fastest iteration cycles on a single product line this year. The piece highlights that both Sunburst and Flare were priced at the same token rate, signaling OpenAI's positioning of image generation as a strategic priority. It notes the September 8 release arrived less than five months after ChatGPT Images 2.0 launched with thinking capabilities, underscoring the rapid cadence. The author attributes the model creations explicitly to OpenAI, preserving correct creator attribution for developers tracking provenance.

OpenAI's official image prompting guide documents GPT Image 2.5 as a two-variant family. "GPT Image 2.5 Flare" is described as the small, speed-optimized model with image quality comparable to GPT Image 2, while "GPT Image 2.5 Sunburst" is positioned as the base, quality-optimized model offering higher image quality th For migration guidance, the guide recommends starting a new workflow with GPT Image 2.5 Flare when speed is the priority or GPT Image 2.5 Sunburst when demanding quality is the priority, then inspecting output and looking for opportunities to reduce latency. Users migrating from an existing, validated GPT Image 2 workf

On September 8, 2026, OpenAI announced ChatGPT Images 2.5 alongside two new Images API model IDs: gpt-image-2.5-flare and gpt-image-2.5-sunburst, replacing the single-model approach of the prior generation. According to a third-party guide based on hands-on API testing (41 generations), bare model IDs like "gpt-image-2 OpenAI's own positioning frames GPT Image 2.5 Sunburst as the precision-first variant (versus Flare for high-throughput workflows), with consumer-side features including Sketch (compositional drawing reference), Templates, comment-scoped image edits, and prompt sharing rolling out across ChatGPT, ChatGPT Work, and Code

The LLMLearner spec sheet for GPT Image 2.5 Sunburst confirms core metadata: an OpenAI multimodal model released September 8, 2026, accepting text and image input and producing image output, with closed/API-only weights and proprietary commercial use. Architecture details remain undisclosed, and the model is served exc The page provides no benchmark numbers, no community ratings, and no Sunburst-specific performance data distinguishing it from the Flare variant. It functions as a thin metadata reference rather than substantive technical coverage, but does correctly identify the model name, provider (OpenAI), modality, and release dat

LLM Stats confirms that GPT Image 2.5 Sunburst was released on September 8, 2026, by OpenAI, with proprietary licensing and multimodal support for both text and image inputs. Pricing starts at $0.0530 per generated image via OpenAI, consistent with the $30 per million output tokens rate from the official documentation. The tracker lists OpenAI as the sole tracked provider and notes the model has not yet accumulated arena votes. The page situates Sunburst among related image models such as GPT Image 2, Grok Imagine Image 2.0, and Gemini 3.1 Flash Image, providing a comparative reference for developers evaluating the ecosystem. It also links to OpenAI's developer documentation and official playground as primary access points. Creator attribution is preserved as OpenAI, with no claim of alternative origins.

OpenAI's developer documentation introduces GPT-Image-2.5 Sunburst as the company's most capable model for image generation and editing, accepting text and image inputs while outputting images. It supports low, medium, high, xhigh, max, and auto quality settings and is selectable via the Image API or the Responses API image generation tool. The page lists text input at $5.00 per million tokens, image input at $8.00 per million tokens, and image output at $30.00 per million tokens with standard processing. Audio and video modalities are not supported on this endpoint. The documentation clarifies that text output is not billed because the model outputs images rather than text, and that image output drops to $15 per million tokens under Batch processing. It also warns that the GPT Image 2 token calculator does not estimate GPT Image 2.5 consumption, an important caveat for developers migrating workloads. The page anchors the model on OpenAI's live v1 endpoints and is positioned as a higher-quality option compared to the faster, medium-priced sibling variant.