{
  "@context": "https://schema.org",
  "@id": "https://aggr.aymericbeaumet.com/items/ampcode-com/2026-06-05-faster-deep-rush/#webpage",
  "@type": [
    "WebPage",
    "ArchiveComponent"
  ],
  "articleBody": "The first token now arrives 87% faster and entire responses are 32% faster, p50, in Amp's `deep` and `rush` modes.\n\nHow? Mostly using websockets for communication with OpenAI, partly because we [rebuilt Amp](https://ampcode.com/news/neo) to be much faster last month.\n\nThese gains matter most on long-horizon tasks, where we're seeing up to a 40% end-to-end speedup from user prompt submission to completion.\n\n![Chart showing OpenAI WebSocket responses finishing faster than HTTP SSE over a long Amp session](https://static.ampcode.com/news/openai-ws-elapsed-by-turn.svg?v=2)\n\nYou can see the difference:\n",
  "authors": [],
  "captured_at": "2026-10-03T23:09:50.514863041+00:00",
  "category": "programming",
  "content_html": "<p>The first token now arrives 87% faster and entire responses are 32% faster, p50, in Amp's <code>deep</code> and <code>rush</code> modes.</p>\n<p>How? Mostly using websockets for communication with OpenAI, partly because we <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https://ampcode.com/news/neo\">rebuilt Amp</a> to be much faster last month.</p>\n<p>These gains matter most on long-horizon tasks, where we're seeing up to a 40% end-to-end speedup from user prompt submission to completion.</p>\n<p><picture class=\"article-picture article-picture-fallback\" style=\"--image-ratio:16 / 9\"><img src=\"https://static.ampcode.com/news/openai-ws-elapsed-by-turn.svg?v=2\" alt=\"Chart showing OpenAI WebSocket responses finishing faster than HTTP SSE over a long Amp session\" loading=\"eager\" decoding=\"async\" fetchpriority=\"high\" referrerpolicy=\"no-referrer\" class=\"progressive-image\"></picture></p>\n<p>You can see the difference:</p>\n",
  "content_markdown": "The first token now arrives 87% faster and entire responses are 32% faster, p50, in Amp's `deep` and `rush` modes.\n\nHow? Mostly using websockets for communication with OpenAI, partly because we [rebuilt Amp](https://ampcode.com/news/neo) to be much faster last month.\n\nThese gains matter most on long-horizon tasks, where we're seeing up to a 40% end-to-end speedup from user prompt submission to completion.\n\n![Chart showing OpenAI WebSocket responses finishing faster than HTTP SSE over a long Amp session](https://static.ampcode.com/news/openai-ws-elapsed-by-turn.svg?v=2)\n\nYou can see the difference:\n",
  "dateCreated": "2026-10-03T23:09:50.514863041+00:00",
  "date_created": "2026-06-05T00:00:00+00:00",
  "date_modified": "2026-06-05T00:00:00+00:00",
  "date_published": "2026-06-05T00:00:00+00:00",
  "description": "The first token now arrives 87% faster and entire responses are 32% faster, p50, in Amp's deep and rush modes. How? Mostly using websockets for communication with OpenAI, partly because we rebuilt Amp to be much faster last month. These…",
  "external_url": "https://ampcode.com/news/faster-deep-rush",
  "git": {
    "commit": "ebb0941c5c06a2ab2377342d607535262a3e969e",
    "path": "items/ampcode-com/2026/06/2026-06-05-faster-deep-rush.md",
    "permalink": "https://github.com/aymericbeaumet/aggr-instance/blob/ebb0941c5c06a2ab2377342d607535262a3e969e/items/ampcode-com/2026/06/2026-06-05-faster-deep-rush.md"
  },
  "id": "urn:aggr:item:350a7b2f3ee68139b12dfffd4c14b09cd7c1b198",
  "isBasedOn": {
    "@id": "https://ampcode.com/news/faster-deep-rush"
  },
  "mainEntity": {
    "@id": "https://ampcode.com/news/faster-deep-rush",
    "@type": "CreativeWork",
    "archivedAt": {
      "@id": "https://aggr.aymericbeaumet.com/items/ampcode-com/2026-06-05-faster-deep-rush/#webpage"
    },
    "dateModified": "2026-06-05T00:00:00+00:00",
    "datePublished": "2026-06-05T00:00:00+00:00",
    "headline": "Faster Deep & Rush",
    "timeRequired": "PT1M",
    "url": "https://ampcode.com/news/faster-deep-rush",
    "wordCount": 69
  },
  "name": "Archived snapshot: Faster Deep & Rush",
  "preview": null,
  "reading_minutes": 1,
  "replicated_at": null,
  "source": "ampcode-com",
  "source_name": "Amp News",
  "summary": "The first token now arrives 87% faster and entire responses are 32% faster, p50, in Amp's deep and rush modes. How? Mostly using websockets for communication with OpenAI, partly because we rebuilt Amp to be much faster last month. These gains matter most on long-horizon tasks, where we're seeing up to a 40% end-to-end speedup from user prompt submission to completion. You can see the difference:",
  "tags": [],
  "temporalCoverage": "2026-10-03T23:09:50.514863041+00:00",
  "title": "Faster Deep & Rush",
  "url": "https://aggr.aymericbeaumet.com/items/ampcode-com/2026-06-05-faster-deep-rush/",
  "word_count": 69
}