๐Ÿฆ Twitter Post Details

Viewing enriched Twitter post

@xywang626

We are super excited to release OpenCUA โ€” the first from 0 to 1 computer-use agent foundation model framework and open-source SOTA model OpenCUA-32B, matching top proprietary models on OSWorld-Verified, with full infrastructure and data. ๐Ÿ”— [Paper] https://t.co/SYEio5ccNJ ๐Ÿ“Œ [Website] https://t.co/ma6bBuYiNM ๐Ÿค– [Models] https://t.co/7TVtIdjkmq ๐Ÿ“Š[Data] https://t.co/N6tQQwQkhs ๐Ÿ’ป [Code] https://t.co/ihr8TXmG6k ๐ŸŒŸ OpenCUA โ€” comprehensive open-source framework for computer-use agents, including: ๐Ÿ“Š AgentNet โ€” first large-scale CUA dataset (3 systems, 200+ apps & sites, 22.6K trajectories) ๐Ÿ† OpenCUA model โ€” open-source SOTA on OSWorld-Verified (34.8% avg success, outperforms OpenAI CUA) ๐Ÿ–ฅ AgentNetTool โ€” cross-system computer-use task annotation tool ๐Ÿ AgentNetBench โ€” offline CUA benchmark for fast, reproducible evaluation ๐Ÿ’ก Why OpenCUA? Proprietary CUAs like Claude or OpenAI CUA are impressive๐Ÿคฏ โ€” but thereโ€™s no large-scale open desktop agent dataset or transparent pipeline. OpenCUA changes that by offering the full open-source stack ๐Ÿ› : scalable cross-system data collection, effective data formulation, model training strategy, and reproducible evaluation โ€” powering top open-source models including OpenCUA-7B and OpenCUA-32B that excel in GUI planning & grounding. Details of OpenCUA framework๐Ÿ‘‡

Media 1
Media 2
Media 3
Media 4

๐Ÿ“Š Media Metadata

{
  "media": [
    {
      "url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/1956400403911962757/media_0.jpg?",
      "media_url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/1956400403911962757/media_0.jpg?",
      "type": "photo",
      "filename": "media_0.jpg"
    },
    {
      "url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/1956400403911962757/media_2.jpg?",
      "media_url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/1956400403911962757/media_2.jpg?",
      "type": "photo",
      "filename": "media_2.jpg"
    },
    {
      "url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/1956400403911962757/media_3.jpg?",
      "media_url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/1956400403911962757/media_3.jpg?",
      "type": "photo",
      "filename": "media_3.jpg"
    },
    {
      "url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/1956400403911962757/media_4.jpg?",
      "media_url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/1956400403911962757/media_4.jpg?",
      "type": "photo",
      "filename": "media_4.jpg"
    }
  ],
  "processed_at": "2025-08-15T22:38:58.998906",
  "pipeline_version": "2.0"
}

๐Ÿ”ง Raw API Response

{
  "type": "tweet",
  "id": "1956400403911962757",
  "url": "https://x.com/xywang626/status/1956400403911962757",
  "twitterUrl": "https://twitter.com/xywang626/status/1956400403911962757",
  "text": "We are super excited to release OpenCUA โ€” the first from 0 to 1 computer-use agent foundation model framework and open-source SOTA model OpenCUA-32B, matching top proprietary models on OSWorld-Verified, with full infrastructure and data.\n\n๐Ÿ”— [Paper] https://t.co/SYEio5ccNJ \n๐Ÿ“Œ [Website] https://t.co/ma6bBuYiNM \n๐Ÿค– [Models] https://t.co/7TVtIdjkmq\n๐Ÿ“Š[Data] https://t.co/N6tQQwQkhs\n๐Ÿ’ป [Code] https://t.co/ihr8TXmG6k\n\n๐ŸŒŸ OpenCUA โ€” comprehensive open-source framework for computer-use agents, including:\n๐Ÿ“Š AgentNet โ€” first large-scale CUA dataset (3 systems, 200+ apps & sites, 22.6K trajectories)\n๐Ÿ† OpenCUA model โ€” open-source SOTA on OSWorld-Verified (34.8% avg success, outperforms OpenAI CUA)\n๐Ÿ–ฅ AgentNetTool โ€” cross-system computer-use task annotation tool\n๐Ÿ AgentNetBench โ€” offline CUA benchmark for fast, reproducible evaluation\n\n๐Ÿ’ก Why OpenCUA?\nProprietary CUAs like Claude or OpenAI CUA are impressive๐Ÿคฏ โ€” but thereโ€™s no large-scale open desktop agent dataset or transparent pipeline. OpenCUA changes that by offering the full open-source stack ๐Ÿ› : scalable cross-system data collection, effective data formulation, model training strategy, and reproducible evaluation โ€” powering top open-source models including OpenCUA-7B and OpenCUA-32B that excel in GUI planning & grounding.\n\nDetails of OpenCUA framework๐Ÿ‘‡",
  "source": "Twitter for iPhone",
  "retweetCount": 33,
  "replyCount": 5,
  "likeCount": 91,
  "quoteCount": 4,
  "viewCount": 12553,
  "createdAt": "Fri Aug 15 16:59:39 +0000 2025",
  "lang": "en",
  "bookmarkCount": 54,
  "isReply": false,
  "inReplyToId": null,
  "conversationId": "1956400403911962757",
  "displayTextRange": [
    0,
    276
  ],
  "inReplyToUserId": null,
  "inReplyToUsername": null,
  "author": {
    "type": "user",
    "userName": "xywang626",
    "url": "https://x.com/xywang626",
    "twitterUrl": "https://twitter.com/xywang626",
    "id": "1645829082121404416",
    "name": "Xinyuan Wang",
    "isVerified": false,
    "isBlueVerified": false,
    "verifiedType": null,
    "profilePicture": "https://pbs.twimg.com/profile_images/1747342445388681216/KQJXFgCw_normal.jpg",
    "coverPicture": "https://pbs.twimg.com/profile_banners/1645829082121404416/1688106058",
    "description": "",
    "location": "Hong Kong",
    "followers": 122,
    "following": 236,
    "status": "",
    "canDm": false,
    "canMediaTag": true,
    "createdAt": "Tue Apr 11 16:40:13 +0000 2023",
    "entities": {
      "description": {
        "urls": []
      },
      "url": {}
    },
    "fastFollowersCount": 0,
    "favouritesCount": 128,
    "hasCustomTimelines": true,
    "isTranslator": false,
    "mediaCount": 15,
    "statusesCount": 44,
    "withheldInCountries": [],
    "affiliatesHighlightedLabel": {},
    "possiblySensitive": false,
    "pinnedTweetIds": [],
    "profile_bio": {
      "description": "PhD Student  @XLangNLP and @hkunlp2020",
      "entities": {
        "description": {
          "user_mentions": [
            {
              "id_str": "0",
              "indices": [
                13,
                22
              ],
              "name": "",
              "screen_name": "XLangNLP"
            },
            {
              "id_str": "0",
              "indices": [
                27,
                38
              ],
              "name": "",
              "screen_name": "hkunlp2020"
            }
          ]
        },
        "url": {
          "urls": [
            {
              "display_url": "xinyuanwangcs.github.io",
              "expanded_url": "https://xinyuanwangcs.github.io/",
              "indices": [
                0,
                23
              ],
              "url": "https://t.co/I8QB9TWxZH"
            }
          ]
        }
      }
    },
    "isAutomated": false,
    "automatedBy": null
  },
  "extendedEntities": {
    "media": [
      {
        "allow_download_status": {
          "allow_download": true
        },
        "display_url": "pic.twitter.com/4LrwHnzWC6",
        "expanded_url": "https://twitter.com/xywang626/status/1956400403911962757/photo/1",
        "ext_media_availability": {
          "status": "Available"
        },
        "features": {
          "large": {},
          "orig": {}
        },
        "id_str": "1956377694347923456",
        "indices": [
          277,
          300
        ],
        "media_key": "3_1956377694347923456",
        "media_results": {
          "id": "QXBpTWVkaWFSZXN1bHRzOgwAAQoAARsmcyXOGxAACgACGyaHzUmacIUAAA==",
          "result": {
            "__typename": "ApiMedia",
            "id": "QXBpTWVkaWE6DAABCgABGyZzJc4bEAAKAAIbJofNSZpwhQAA",
            "media_key": "3_1956377694347923456"
          }
        },
        "media_url_https": "https://pbs.twimg.com/media/GyZzJc4bEAATQap.jpg",
        "original_info": {
          "focus_rects": [
            {
              "h": 1454,
              "w": 2596,
              "x": 843,
              "y": 0
            },
            {
              "h": 1454,
              "w": 1454,
              "x": 1594,
              "y": 0
            },
            {
              "h": 1454,
              "w": 1275,
              "x": 1684,
              "y": 0
            },
            {
              "h": 1454,
              "w": 727,
              "x": 1958,
              "y": 0
            },
            {
              "h": 1454,
              "w": 3439,
              "x": 0,
              "y": 0
            }
          ],
          "height": 1454,
          "width": 3439
        },
        "sizes": {
          "large": {
            "h": 866,
            "w": 2048
          }
        },
        "type": "photo",
        "url": "https://t.co/4LrwHnzWC6"
      }
    ]
  },
  "card": null,
  "place": {},
  "entities": {
    "urls": [
      {
        "display_url": "arxiv.org/abs/2508.09123",
        "expanded_url": "https://arxiv.org/abs/2508.09123",
        "indices": [
          249,
          272
        ],
        "url": "https://t.co/SYEio5ccNJ"
      },
      {
        "display_url": "opencua.xlang.ai",
        "expanded_url": "https://opencua.xlang.ai/",
        "indices": [
          286,
          309
        ],
        "url": "https://t.co/ma6bBuYiNM"
      },
      {
        "display_url": "huggingface.co/xlangai/OpenCUโ€ฆ",
        "expanded_url": "http://huggingface.co/xlangai/OpenCUA-32B",
        "indices": [
          322,
          345
        ],
        "url": "https://t.co/7TVtIdjkmq"
      },
      {
        "display_url": "huggingface.co/datasets/xlangโ€ฆ",
        "expanded_url": "http://huggingface.co/datasets/xlangai/AgentNet",
        "indices": [
          354,
          377
        ],
        "url": "https://t.co/N6tQQwQkhs"
      },
      {
        "display_url": "github.com/xlang-ai/OpenCโ€ฆ",
        "expanded_url": "https://github.com/xlang-ai/OpenCUA",
        "indices": [
          387,
          410
        ],
        "url": "https://t.co/ihr8TXmG6k"
      }
    ]
  },
  "quoted_tweet": null,
  "retweeted_tweet": null,
  "article": null
}