- 🆕 OpenAI image generation model in the GPT Image 2.5 line, released September 2026.
- ⚡ Positioned by OpenAI for fast, high-quality everyday image generation.
- 🎯 Sibling Sunburst is recommended instead when editing precision matters most.
- 🔧 Callable directly through the Image API by model name.
- 💬 Also selectable inside the Responses API image generation tool.
- 👁️ Accepts image inputs, enabling conversational generate-or-edit workflows.
- 🏢 Follows the documented GPT Image 2 and GPT Image 1.5 models.
OpenAI is an American artificial intelligence research organization headquartered in San Francisco, structured as both a for-profit public benefit corporation and a nonprofit foundation. The lab developed the GPT family of large language models, the DALL-E image generation…
Explore 27 more models by OpenAI →GPT Image 2.5 Flare is one of two image models in OpenAI's GPT Image 2.5 release, arriving alongside GPT Image 2.5 Sunburst. In OpenAI's own guidance, the two are split by workload: Flare is described as the choice for fast, high-quality everyday image generation, while Sunburst is recommended for workflows where editing precision matters most. An inpainting-oriented variant, GPT Image 2.5 Flare Edit, covers masked edit requests in this catalog.
Compared with its same-family predecessors, the main change is this explicit specialization. GPT Image 1.5 was documented primarily as an improvement in instruction following and prompt adherence, and GPT Image 2 as a single model handling both fast, high-quality generation and editing, with flexible image sizes and high-fidelity image inputs. With 2.5, OpenAI instead ships a generation-leaning model and an editing-leaning model, letting developers pick per task rather than tuning one endpoint for everything.
Integration also broadens. Flare can be requested directly from the Image API for one-shot text-to-image or edit calls, or nominated inside the image generation tool of the Responses API, where a mainline text model such as GPT-6 Astra orchestrates the request. In that mode an action parameter decides whether the model creates a new image or edits one already in context, supporting multi-turn, conversational image editing.
OpenAI has not published parameter counts, architecture details, or license terms for the GPT Image family, so it should be treated as a closed, API-served model.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
| Seller | Reputation↓ | Routing | $ / img | Categories | API |
|---|---|---|---|---|---|
| ▲ Apex Ant 0x73b4…e736 | 91.01 | #1 | $0.12 | image,media | openai-images |
"Best price" and the seller table are live AntSeed catalog data (advertised $/1M tokens — or $ per generated image for unit-billed image models — not settled amounts). Reputation = on-chain trust (0-100). "Routing" = the SDK's default buyer routing (what the VPR desktop app ships with): a trust ≥ 60 gate on the effective reputation, then cheapest-first among routable sellers; live failover state (per-peer cooldowns) is buyer-side runtime and not included. Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase. "Usage on AntSeed" counts only settlements whose buyers share the per-model split on-chain (metadata v2/v3, opt-in), so every usage figure is a lower bound.