{
  "@context": "https://schema.org",
  "@id": "https://aggr.aymericbeaumet.com/items/ampcode-com/2026-03-26-gpt-5-4-in-deep/#webpage",
  "@type": [
    "WebPage",
    "ArchiveComponent"
  ],
  "articleBody": "GPT-5.4 now powers Amp's [deep mode](https://ampcode.com/news/deep-mode).\n\nIt's the best model in the world right now.\n\nIt's faster than GPT-5.3-Codex and still as great at coding.\n\nBut, out of the box, GPT-5.4 was too chatty. That's not what we want for `deep`; it's not a pair programmer, it's supposed to go off and solve the problem.\n\nSo we tuned GPT-5.4 to behave like GPT-5.3-Codex.\n\nOnce we had that, we started to use it exclusively; even for interactive tasks. We run it at very high reasoning (`deep^3`) and still prefer it when we need fast interaction and fast reaction. It takes steering better than GPT-5.3-Codex.\n\nTo use it: open the command palette with `Ctrl-O` and run the command `mode: use deep` in the Amp CLI, or select `deep` mode in the Amp editor extension's prompt field. By default it uses `deep^2`, you can switch to `deep^3` by hitting `Opt-D`.\n",
  "authors": [],
  "captured_at": "2026-10-03T23:09:50.514863041+00:00",
  "category": "programming",
  "content_html": "<p>GPT-5.4 now powers Amp's <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https://ampcode.com/news/deep-mode\">deep mode</a>.</p>\n<p>It's the best model in the world right now.</p>\n<p>It's faster than GPT-5.3-Codex and still as great at coding.</p>\n<p>But, out of the box, GPT-5.4 was too chatty. That's not what we want for <code>deep</code>; it's not a pair programmer, it's supposed to go off and solve the problem.</p>\n<p>So we tuned GPT-5.4 to behave like GPT-5.3-Codex.</p>\n<p>Once we had that, we started to use it exclusively; even for interactive tasks. We run it at very high reasoning (<code>deep^3</code>) and still prefer it when we need fast interaction and fast reaction. It takes steering better than GPT-5.3-Codex.</p>\n<p>To use it: open the command palette with <code>Ctrl-O</code> and run the command <code>mode: use deep</code> in the Amp CLI, or select <code>deep</code> mode in the Amp editor extension's prompt field. By default it uses <code>deep^2</code>, you can switch to <code>deep^3</code> by hitting <code>Opt-D</code>.</p>\n",
  "content_markdown": "GPT-5.4 now powers Amp's [deep mode](https://ampcode.com/news/deep-mode).\n\nIt's the best model in the world right now.\n\nIt's faster than GPT-5.3-Codex and still as great at coding.\n\nBut, out of the box, GPT-5.4 was too chatty. That's not what we want for `deep`; it's not a pair programmer, it's supposed to go off and solve the problem.\n\nSo we tuned GPT-5.4 to behave like GPT-5.3-Codex.\n\nOnce we had that, we started to use it exclusively; even for interactive tasks. We run it at very high reasoning (`deep^3`) and still prefer it when we need fast interaction and fast reaction. It takes steering better than GPT-5.3-Codex.\n\nTo use it: open the command palette with `Ctrl-O` and run the command `mode: use deep` in the Amp CLI, or select `deep` mode in the Amp editor extension's prompt field. By default it uses `deep^2`, you can switch to `deep^3` by hitting `Opt-D`.\n",
  "dateCreated": "2026-10-03T23:09:50.514863041+00:00",
  "date_created": "2026-03-26T00:00:00+00:00",
  "date_modified": "2026-03-26T00:00:00+00:00",
  "date_published": "2026-03-26T00:00:00+00:00",
  "description": "GPT-5.4 now powers Amp's deep mode. It's the best model in the world right now. It's faster than GPT-5.3-Codex and still as great at coding. But, out of the box, GPT-5.4 was too chatty. That's not what we want for deep; it's not a pair…",
  "external_url": "https://ampcode.com/news/gpt-5.4",
  "git": {
    "commit": "ebb0941c5c06a2ab2377342d607535262a3e969e",
    "path": "items/ampcode-com/2026/03/2026-03-26-gpt-5-4-in-deep.md",
    "permalink": "https://github.com/aymericbeaumet/aggr-instance/blob/ebb0941c5c06a2ab2377342d607535262a3e969e/items/ampcode-com/2026/03/2026-03-26-gpt-5-4-in-deep.md"
  },
  "id": "urn:aggr:item:595e8ade533133a7bf198f96eed737ed59731bfc",
  "isBasedOn": {
    "@id": "https://ampcode.com/news/gpt-5.4"
  },
  "mainEntity": {
    "@id": "https://ampcode.com/news/gpt-5.4",
    "@type": "CreativeWork",
    "archivedAt": {
      "@id": "https://aggr.aymericbeaumet.com/items/ampcode-com/2026-03-26-gpt-5-4-in-deep/#webpage"
    },
    "dateModified": "2026-03-26T00:00:00+00:00",
    "datePublished": "2026-03-26T00:00:00+00:00",
    "headline": "GPT‐5.4 in Deep",
    "timeRequired": "PT1M",
    "url": "https://ampcode.com/news/gpt-5.4",
    "wordCount": 161
  },
  "name": "Archived snapshot: GPT‐5.4 in Deep",
  "preview": null,
  "reading_minutes": 1,
  "replicated_at": null,
  "source": "ampcode-com",
  "source_name": "Amp News",
  "summary": "GPT-5.4 now powers Amp's deep mode. It's the best model in the world right now. It's faster than GPT-5.3-Codex and still as great at coding. But, out of the box, GPT-5.4 was too chatty. That's not what we want for deep; it's not a pair programmer, it's supposed to go off and solve the problem. So we tuned GPT-5.4 to behave like GPT-5.3-Codex. Once we had that, we started to use it exclusively; even for interactive tasks. We run it at very high reasoning (deep^3) and still prefer it when we need fast interaction and fast reaction. It takes steering better than GPT-5.3-Codex. To use it: open the command palette with Ctrl-O and run the command mode: use deep in the Amp CLI, or select deep mode in the Amp editor extension's prompt field. By default it uses deep^2, you can switch to deep^3 by hitting Opt-D.",
  "tags": [],
  "temporalCoverage": "2026-10-03T23:09:50.514863041+00:00",
  "title": "GPT‐5.4 in Deep",
  "url": "https://aggr.aymericbeaumet.com/items/ampcode-com/2026-03-26-gpt-5-4-in-deep/",
  "word_count": 161
}