GPT Image 2.6 Flare
openai/gpt-image-2.6-flareGPT Image 2.6 Flare是OpenAI推出的图像生成与编辑模型,针对速度进行了优化:支持快速迭代以及交互式文生图、图生图工作流,生成速度比GPT Image 2.5快约50%。该模型接收文本和图像输入并生成图像,画质档位从低到最高可选,支持透明背景,可自定义尺寸,尺寸需为16像素的倍数,宽高比范围为1:3至3:1。GPT Image 2.6新增复杂度设置,用于控制模型在渲染前的规划程度,同时支持标记引导编辑、动态GIF输出以及SVG输出。
- Modalities
- Public starting price
- Priced by multiple attributes
- Added to catalog
- Nov 10, 2026
Online Inference
Configure input
Controls and defaults come directly from this model’s API schema.
Task result

{
"image_urls": [
"https://oss.shengsuanyun.com/example/modelinfo/1493/workbench/8bc3c94a-cd6a-4e7f-bd9b-d85493cbdd6d/output_0.gif"
]
}
Task History
Task history
Only generation records from the last seven days are available.
Sign in to view your task history.
Sign in to view task historyFilled when you open this page. The default query uses the current model ID; leave blank for records from all models.
| Task ID | Model | Completion status | Created | Duration | Credits consumed | Artifact |
|---|
API Call
GPT Image 2.6 Flare code examples and API guide
Modelmesh normalizes requests and responses across service providers behind one consistent API.
Modelmesh provides an OpenAI-compatible Completion API for more than 300 models and service providers. Call it directly, through the OpenAI SDK, or with supported third-party SDKs.
Modelmesh-specific request headers in these examples are optional. When supplied, your application can appear on the Modelmesh rankings.
#!/usr/bin/env bash
set -euo pipefail
API_KEY="$SSY_API_KEY"
curl --fail-with-body --silent --show-error --location \
--request POST \
"https://router.shengsuanyun.com/api/v1/tasks/generations" \
--header "Content-Type: application/json" \
--header "Authorization: Bearer ${API_KEY}" \
--data '{
"model": "openai/gpt-image-2.6-flare",
"background": "opaque",
"frame_count": 16,
"frame_duration_ms": 100,
"images": [
{
"image_url": "https://oss.shengsuanyun.com/example/modelinfo/1493/workbench/8bc3c94a-cd6a-4e7f-bd9b-d85493cbdd6d/input_0.jpg"
}
],
"moderation": "auto",
"n": 1,
"output_format": "gif",
"prompt": "以提供的参考图片为基础,制作一张自然、精致、可无缝循环播放的 GIF。\n\n【保留参考图】\n准确保留参考图片中主体的身份、外形、比例、颜色、材质及构图。保留原有背景、光照方向,以及已有文字和标志的内容与位置。不新增物体,不改变产品结构,不重新设计画面。\n\n【动画设计】\n让主体呈现非常轻微、平滑的上下悬浮运动,位移不超过画面高度的 2%,完成一次缓慢上浮再回到原位的完整周期。如果画面中存在主体的投影,让投影的柔和程度与大小随悬浮同步轻微变化。主体表面的高光随运动自然变化,避免突兀的闪光。\n\n【时间与连续性】\n动画从参考图的原始位置开始,平滑加速、减速,并在周期结束时回到起始状态。首尾位置、速度、光照与阴影连续,形成自然的无缝循环。所有帧保持主体细节稳定,不能出现形变、纹理漂移、文字变化、闪烁或额外肢体。\n\n【镜头】\n镜头完全固定,不推拉、不旋转、不切换场景。背景保持静止,主体运动克制,整体呈现高品质的动态产品展示效果。",
"quality": "high",
"size": "auto",
"stream": false
}'
#!/usr/bin/env bash
set -euo pipefail
API_KEY="$SSY_API_KEY"
REQUEST_ID="${SSY_REQUEST_ID:?Set SSY_REQUEST_ID to the ID returned by task submission}"
curl --fail-with-body --silent --show-error --location \
--request GET \
"https://router.shengsuanyun.com/api/v1/tasks/generations/${REQUEST_ID}" \
--header "Authorization: Bearer ${API_KEY}"
Pricing
Pricing Details
All prices shown in CNY- 图片·输入
- ¥56/M token
- 图片·输出
- ¥210/M token
- 图片缓存读
- ¥14/M token
- 缓存写(5m)
- ¥70/M token
- Text · Input
- ¥35/M token
- Text · Output
- ¥210/M token