{
  "@context": "https://schema.org",
  "@id": "https://aggr.aymericbeaumet.com/items/simonwillison-net/2026-09-08-introducing-chatgpt-images-2-5/#webpage",
  "@type": [
    "WebPage",
    "ArchiveComponent"
  ],
  "articleBody": "**[Introducing ChatGPT Images 2.5](https://openai.com/index/introducing-chatgpt-images-2-5/)**. OpenAI's image generation models are apparently used \"more than 3 billion images across ChatGPT Images and the GPT‑Image models in the API\". This latest release improves their instruction-following ability across multiple turns, responds faster, and \"is better at preserving the subjects in your reference photos\".\n\nThere are two new model IDs in the API: `gpt-image-2.5-sunburst` and `gpt-image-2.5-flare`. Based [on this](https://developers.openai.com/api/docs/guides/image-generation#overview) I think Sunburst is the stronger option:\n\n> Choose Sunburst for workflows where editing precision matters most, and Flare for fast, high-quality everyday image generation.\n\nI [upgraded](https://github.com/simonw/tools/pull/333) my [openai\\_image.py](https://tools.simonwillison.net/python/#openai_imagepy) CLI tool to support passing in one or more reference images, so now this works:\n\n```\nuv run https://tools.simonwillison.net/python/openai_image.py \\\n  'add a raccoon scientist studying the chart thoughtfully' \\\n  -i https://static.simonwillison.net/static/2026/openai-agent-usage.webp \\\n  -m gpt-image-2.5-sunburst\n```\n\nThis is the [original image](https://static.simonwillison.net/static/2026/openai-agent-usage.webp), and here's what I got back from that prompt to \"add a raccoon scientist studying the chart thoughtfully\":\n\n![Line chart with cartoon illustration. Title: \\\"Usage of internal coding agents is increasing significantly—Median researcher\\\". Y-axis labeled \\\"Daily $ / researcher\\\" from 0 to 700; x-axis shows Feb 2026, Apr 2026, Jun 2026, Aug 2026. A blue line stays near zero through April, rises gradually to around 150 by June and July, then climbs steeply to about 600 by late August. In the foreground a cartoon raccoon in glasses and a white lab coat, chin in hand, holds a clipboard at a desk with a mug bearing the OpenAI logo, some printed charts, and a stack of three books titled \\\"AI AGENTS\\\", \\\"SOFTWARE ENGINEERING\\\", and \\\"PRODUCTIVITY\\\". An OpenAI logo appears in the top right corner.](https://static.simonwillison.net/static/2026/racoon-chart.webp)\n",
  "authors": [],
  "captured_at": "2026-09-08T22:56:52.093968329+00:00",
  "category": "programming",
  "content_html": "<p><strong><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https://openai.com/index/introducing-chatgpt-images-2-5/\">Introducing ChatGPT Images 2.5</a></strong>. OpenAI's image generation models are apparently used &quot;more than 3 billion images across ChatGPT Images and the GPT‑Image models in the API&quot;. This latest release improves their instruction-following ability across multiple turns, responds faster, and &quot;is better at preserving the subjects in your reference photos&quot;.</p>\n<p>There are two new model IDs in the API: <code>gpt-image-2.5-sunburst</code> and <code>gpt-image-2.5-flare</code>. Based <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https://developers.openai.com/api/docs/guides/image-generation#overview\">on this</a> I think Sunburst is the stronger option:</p>\n<blockquote>\n<p>Choose Sunburst for workflows where editing precision matters most, and Flare for fast, high-quality everyday image generation.</p>\n</blockquote>\n<p>I <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https://github.com/simonw/tools/pull/333\">upgraded</a> my <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https://tools.simonwillison.net/python/#openai_imagepy\">openai_image.py</a> CLI tool to support passing in one or more reference images, so now this works:</p>\n<pre><code><span class=\"code-snippet\">uv run https://tools.simonwillison.net/python/openai_image.py \\\n  &#39;add a raccoon scientist studying the chart thoughtfully&#39; \\\n  -i https://static.simonwillison.net/static/2026/openai-agent-usage.webp \\\n  -m gpt-image-2.5-sunburst\n</span></code></pre>\n<p>This is the <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https://static.simonwillison.net/static/2026/openai-agent-usage.webp\">original image</a>, and here's what I got back from that prompt to &quot;add a raccoon scientist studying the chart thoughtfully&quot;:</p>\n<p><picture class=\"article-picture article-picture-fallback\" style=\"--image-ratio:16 / 9\"><img src=\"https://static.simonwillison.net/static/2026/racoon-chart.webp\" alt=\"Line chart with cartoon illustration. Title: &quot;Usage of internal coding agents is increasing significantly—Median researcher&quot;. Y-axis labeled &quot;Daily $ / researcher&quot; from 0 to 700; x-axis shows Feb 2026, Apr 2026, Jun 2026, Aug 2026. A blue line stays near zero through April, rises gradually to around 150 by June and July, then climbs steeply to about 600 by late August. In the foreground a cartoon raccoon in glasses and a white lab coat, chin in hand, holds a clipboard at a desk with a mug bearing the OpenAI logo, some printed charts, and a stack of three books titled &quot;AI AGENTS&quot;, &quot;SOFTWARE ENGINEERING&quot;, and &quot;PRODUCTIVITY&quot;. An OpenAI logo appears in the top right corner.\" loading=\"eager\" decoding=\"async\" fetchpriority=\"high\" referrerpolicy=\"no-referrer\" class=\"progressive-image\"></picture></p>\n",
  "content_markdown": "**[Introducing ChatGPT Images 2.5](https://openai.com/index/introducing-chatgpt-images-2-5/)**. OpenAI's image generation models are apparently used \"more than 3 billion images across ChatGPT Images and the GPT‑Image models in the API\". This latest release improves their instruction-following ability across multiple turns, responds faster, and \"is better at preserving the subjects in your reference photos\".\n\nThere are two new model IDs in the API: `gpt-image-2.5-sunburst` and `gpt-image-2.5-flare`. Based [on this](https://developers.openai.com/api/docs/guides/image-generation#overview) I think Sunburst is the stronger option:\n\n> Choose Sunburst for workflows where editing precision matters most, and Flare for fast, high-quality everyday image generation.\n\nI [upgraded](https://github.com/simonw/tools/pull/333) my [openai\\_image.py](https://tools.simonwillison.net/python/#openai_imagepy) CLI tool to support passing in one or more reference images, so now this works:\n\n```\nuv run https://tools.simonwillison.net/python/openai_image.py \\\n  'add a raccoon scientist studying the chart thoughtfully' \\\n  -i https://static.simonwillison.net/static/2026/openai-agent-usage.webp \\\n  -m gpt-image-2.5-sunburst\n```\n\nThis is the [original image](https://static.simonwillison.net/static/2026/openai-agent-usage.webp), and here's what I got back from that prompt to \"add a raccoon scientist studying the chart thoughtfully\":\n\n![Line chart with cartoon illustration. Title: \\\"Usage of internal coding agents is increasing significantly—Median researcher\\\". Y-axis labeled \\\"Daily $ / researcher\\\" from 0 to 700; x-axis shows Feb 2026, Apr 2026, Jun 2026, Aug 2026. A blue line stays near zero through April, rises gradually to around 150 by June and July, then climbs steeply to about 600 by late August. In the foreground a cartoon raccoon in glasses and a white lab coat, chin in hand, holds a clipboard at a desk with a mug bearing the OpenAI logo, some printed charts, and a stack of three books titled \\\"AI AGENTS\\\", \\\"SOFTWARE ENGINEERING\\\", and \\\"PRODUCTIVITY\\\". An OpenAI logo appears in the top right corner.](https://static.simonwillison.net/static/2026/racoon-chart.webp)\n",
  "dateCreated": "2026-09-08T22:56:52.093968329+00:00",
  "date_created": "2026-09-08T22:46:33+00:00",
  "date_modified": "2026-09-08T22:46:33+00:00",
  "date_published": "2026-09-08T22:46:33+00:00",
  "description": "Introducing ChatGPT Images 2.5 OpenAI's image generation models are apparently used \"more than 3 billion images across ChatGPT Images and the GPT‑Image models in the API\". This latest release improves their instruction-following ability…",
  "external_url": "https://simonwillison.net/2026/Sep/8/introducing-chatgpt-images-25/",
  "git": {
    "commit": "ebb0941c5c06a2ab2377342d607535262a3e969e",
    "path": "items/simonwillison-net/2026/09/2026-09-08-introducing-chatgpt-images-2-5.md",
    "permalink": "https://github.com/aymericbeaumet/aggr-instance/blob/ebb0941c5c06a2ab2377342d607535262a3e969e/items/simonwillison-net/2026/09/2026-09-08-introducing-chatgpt-images-2-5.md"
  },
  "id": "urn:aggr:item:e4dfa15195675d509474ee51b2caf5a06d465acd",
  "isBasedOn": {
    "@id": "https://simonwillison.net/2026/Sep/8/introducing-chatgpt-images-25/"
  },
  "mainEntity": {
    "@id": "https://simonwillison.net/2026/Sep/8/introducing-chatgpt-images-25/",
    "@type": "CreativeWork",
    "archivedAt": {
      "@id": "https://aggr.aymericbeaumet.com/items/simonwillison-net/2026-09-08-introducing-chatgpt-images-2-5/#webpage"
    },
    "dateModified": "2026-09-08T22:46:33+00:00",
    "datePublished": "2026-09-08T22:46:33+00:00",
    "headline": "Introducing ChatGPT Images 2.5",
    "timeRequired": "PT1M",
    "url": "https://simonwillison.net/2026/Sep/8/introducing-chatgpt-images-25/",
    "wordCount": 166
  },
  "name": "Archived snapshot: Introducing ChatGPT Images 2.5",
  "preview": null,
  "reading_minutes": 1,
  "replicated_at": null,
  "source": "simonwillison-net",
  "source_name": "Simon Willison's Weblog",
  "summary": "Introducing ChatGPT Images 2.5 OpenAI's image generation models are apparently used \"more than 3 billion images across ChatGPT Images and the GPT‑Image models in the API\". This latest release improves their instruction-following ability across multiple turns, responds faster, and \"is better at preserving the subjects in your reference photos\". There are two new model IDs in the API: gpt-image-2.5-sunburst and gpt-image-2.5-flare. Based on this I think Sunburst is the stronger option: Choose Sunburst for workflows where editing precision matters most, and Flare for fast, high-quality everyday image generation. I upgraded my openai_image.py CLI tool to support passing in one or more reference images, so now this works: uv run https://tools.simonwillison.net/python/openai_image.py \\ 'add a raccoon scientist studying the chart thoughtfully' \\ -i https://static.simonwillison.net/static/2026/openai-agent-usage.webp \\ -m gpt-image-2.5-sunburst This is the original image, and here's what I got back from that prompt to \"add a raccoon scientist studying the chart thoughtfully\": Tags: tools, ai, openai, generative-ai, uv, text-to-image",
  "tags": [
    "ai",
    "generative-ai",
    "openai",
    "text-to-image",
    "tools",
    "uv"
  ],
  "temporalCoverage": "2026-09-08T22:56:52.093968329+00:00",
  "title": "Introducing ChatGPT Images 2.5",
  "url": "https://aggr.aymericbeaumet.com/items/simonwillison-net/2026-09-08-introducing-chatgpt-images-2-5/",
  "word_count": 166
}