🐦 Twitter Post Details

Viewing enriched Twitter post

@BoWang87

Two major AI releases this week: • Qwen3.5 — new open-source small models • GPT-5.4 — newest frontier closed model Most benchmarks compare math and coding. But the real test for frontier AI should be biology and healthcare. That’s where mistakes actually matter. So our team at @UHN ran them on EURORAD — 207 expert-validated radiology differential diagnosis cases. Results: GPT-5.4: 92.2% Qwen3.5-27B: 85% Gemini 3.1 Pro: ~79% A 27B open model that runs on a laptop is only 7 points behind the most powerful AI model on earth — and already beating Gemini on this benchmark. That gap is much smaller than people expected. And it matters. For years hospitals faced an impossible tradeoff: Frontier models → patient data leaves the hospital Local models → not good enough That tradeoff may finally be ending. Qwen3.5-27B runs fully local. No API. No cloud. No patient data leaving the building. HIPAA / PHIPA compliance becomes architecture, not paperwork. Interesting detail: 27B and 122B score almost identically here. Scaling bigger didn’t help much. One caveat: with web-scale training, it’s hard to completely rule out that frontier models like GPT-5.4 may have seen parts of evaluation datasets. Still, the signal is clear: Small models are getting good enough for real clinical AI. And if we want to measure real AI progress, biology and healthcare should be the benchmark. Huge credit to the team @alifmunim @AlhusainAbdalla @JunMa_AI4Health @Omar_Ibr12 @oliviaamwei

Media 1
View on Twitter

📊 Media Metadata

{
  "media": [
    {
      "url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2029649149378072654/media_0.jpg?",
      "media_url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2029649149378072654/media_0.jpg?",
      "type": "photo",
      "filename": "media_0.jpg"
    }
  ],
  "processed_at": "2026-03-06T14:04:29.070638",
  "pipeline_version": "2.0"
}

🔧 Raw API Response

{
  "type": "tweet",
  "id": "2029649149378072654",
  "url": "https://x.com/BoWang87/status/2029649149378072654",
  "twitterUrl": "https://twitter.com/BoWang87/status/2029649149378072654",
  "text": "Two major AI releases this week:\n\n• Qwen3.5 — new open-source small models\n• GPT-5.4 — newest frontier closed model\n\nMost benchmarks compare math and coding.\n\nBut the real test for frontier AI should be biology and healthcare.\n\nThat’s where mistakes actually matter.\n\nSo our team at @UHN ran them on EURORAD — 207 expert-validated radiology differential diagnosis cases.\n\nResults:\n\nGPT-5.4: 92.2%\nQwen3.5-27B: 85%\nGemini 3.1 Pro: ~79%\n\nA 27B open model that runs on a laptop is only 7 points behind the most powerful AI model on earth — and already beating Gemini on this benchmark.\n\nThat gap is much smaller than people expected.\n\nAnd it matters.\n\nFor years hospitals faced an impossible tradeoff:\n\nFrontier models → patient data leaves the hospital\nLocal models → not good enough\n\nThat tradeoff may finally be ending.\n\nQwen3.5-27B runs fully local.\nNo API. No cloud. No patient data leaving the building.\n\nHIPAA / PHIPA compliance becomes architecture, not paperwork.\n\nInteresting detail: 27B and 122B score almost identically here.\nScaling bigger didn’t help much.\n\nOne caveat: with web-scale training, it’s hard to completely rule out that frontier models like GPT-5.4 may have seen parts of evaluation datasets.\n\nStill, the signal is clear:\n\nSmall models are getting good enough for real clinical AI.\n\nAnd if we want to measure real AI progress,\nbiology and healthcare should be the benchmark. \n\nHuge credit to the team\n@alifmunim @AlhusainAbdalla @JunMa_AI4Health @Omar_Ibr12 @oliviaamwei",
  "source": "Twitter for iPhone",
  "retweetCount": 30,
  "replyCount": 24,
  "likeCount": 222,
  "quoteCount": 8,
  "viewCount": 55487,
  "createdAt": "Thu Mar 05 20:04:00 +0000 2026",
  "lang": "en",
  "bookmarkCount": 66,
  "isReply": false,
  "inReplyToId": null,
  "conversationId": "2029649149378072654",
  "displayTextRange": [
    0,
    274
  ],
  "inReplyToUserId": null,
  "inReplyToUsername": null,
  "author": {
    "type": "user",
    "userName": "BoWang87",
    "url": "https://x.com/BoWang87",
    "twitterUrl": "https://twitter.com/BoWang87",
    "id": "775759968221868032",
    "name": "Bo Wang",
    "isVerified": false,
    "isBlueVerified": true,
    "verifiedType": null,
    "profilePicture": "https://pbs.twimg.com/profile_images/1707747805341552640/EPsdigbV_normal.jpg",
    "coverPicture": "https://pbs.twimg.com/profile_banners/775759968221868032/1731284237",
    "description": "",
    "location": "Toronto, Ontario",
    "followers": 24010,
    "following": 2877,
    "status": "",
    "canDm": false,
    "canMediaTag": true,
    "createdAt": "Tue Sep 13 18:16:04 +0000 2016",
    "entities": {
      "description": {
        "urls": []
      },
      "url": {}
    },
    "fastFollowersCount": 0,
    "favouritesCount": 70997,
    "hasCustomTimelines": true,
    "isTranslator": false,
    "mediaCount": 263,
    "statusesCount": 1626,
    "withheldInCountries": [],
    "affiliatesHighlightedLabel": {},
    "possiblySensitive": false,
    "pinnedTweetIds": [
      "2029649149378072654"
    ],
    "profile_bio": {
      "description": "SVP @Xaira_Thera | Assoc Prof @UofT | Building first Virtual Cell |AI & Bio & Healthcare | Inventor of scGPT, MedSAM, BioReason | Opinions my own",
      "entities": {
        "description": {
          "hashtags": [],
          "symbols": [],
          "urls": [],
          "user_mentions": [
            {
              "id_str": "0",
              "indices": [
                4,
                16
              ],
              "name": "",
              "screen_name": "Xaira_Thera"
            },
            {
              "id_str": "0",
              "indices": [
                30,
                35
              ],
              "name": "",
              "screen_name": "UofT"
            }
          ]
        },
        "url": {
          "urls": [
            {
              "display_url": "scholar.google.ca/citations?user…",
              "expanded_url": "https://scholar.google.ca/citations?user=37FDILIAAAAJ&hl=en",
              "indices": [
                0,
                23
              ],
              "url": "https://t.co/KPX1esmsWc"
            }
          ]
        }
      }
    },
    "isAutomated": false,
    "automatedBy": null
  },
  "extendedEntities": {
    "media": [
      {
        "allow_download_status": {
          "allow_download": true
        },
        "display_url": "pic.twitter.com/ZCn8suirrq",
        "expanded_url": "https://twitter.com/BoWang87/status/2029649149378072654/photo/1",
        "ext_media_availability": {
          "status": "Available"
        },
        "features": {
          "large": {
            "faces": [
              {
                "h": 50,
                "w": 50,
                "x": 446,
                "y": 126
              }
            ]
          },
          "orig": {
            "faces": [
              {
                "h": 50,
                "w": 50,
                "x": 446,
                "y": 126
              }
            ]
          }
        },
        "id_str": "2029649146609815553",
        "indices": [
          275,
          298
        ],
        "media_key": "3_2029649146609815553",
        "media_results": {
          "id": "QXBpTWVkaWFSZXN1bHRzOgwAAQoAARwqwyUyWlABCgACHCrDJddakE4AAA==",
          "result": {
            "__typename": "ApiMedia",
            "id": "QXBpTWVkaWE6DAABCgABHCrDJTJaUAEKAAIcKsMl11qQTgAA",
            "media_key": "3_2029649146609815553"
          }
        },
        "media_url_https": "https://pbs.twimg.com/media/HCrDJTJaUAEbdIQ.jpg",
        "original_info": {
          "focus_rects": [
            {
              "h": 781,
              "w": 1394,
              "x": 0,
              "y": 0
            },
            {
              "h": 1292,
              "w": 1292,
              "x": 51,
              "y": 0
            },
            {
              "h": 1292,
              "w": 1133,
              "x": 131,
              "y": 0
            },
            {
              "h": 1292,
              "w": 646,
              "x": 374,
              "y": 0
            },
            {
              "h": 1292,
              "w": 1394,
              "x": 0,
              "y": 0
            }
          ],
          "height": 1292,
          "width": 1394
        },
        "sizes": {
          "large": {
            "h": 1292,
            "w": 1394
          }
        },
        "type": "photo",
        "url": "https://t.co/ZCn8suirrq"
      }
    ]
  },
  "card": null,
  "place": {},
  "entities": {
    "hashtags": [],
    "symbols": [],
    "timestamps": [],
    "urls": [],
    "user_mentions": [
      {
        "id_str": "153085446",
        "indices": [
          283,
          287
        ],
        "name": "University Health Network",
        "screen_name": "UHN"
      },
      {
        "id_str": "916184957885788165",
        "indices": [
          1425,
          1435
        ],
        "name": "Alif Munim (d/acc)",
        "screen_name": "alifmunim"
      },
      {
        "id_str": "2012244020635054080",
        "indices": [
          1436,
          1452
        ],
        "name": "Alhusain Abdalla",
        "screen_name": "AlhusainAbdalla"
      },
      {
        "id_str": "2004981853653204992",
        "indices": [
          1453,
          1469
        ],
        "name": "JunMa_AI4Health",
        "screen_name": "JunMa_AI4Health"
      },
      {
        "id_str": "1856715502657347584",
        "indices": [
          1470,
          1481
        ],
        "name": "Omar Ibrahim",
        "screen_name": "Omar_Ibr12"
      },
      {
        "id_str": "2029251887824797696",
        "indices": [
          1482,
          1494
        ],
        "name": "Meng Wei",
        "screen_name": "oliviaamwei"
      }
    ]
  },
  "quoted_tweet": null,
  "retweeted_tweet": null,
  "isLimitedReply": false,
  "article": null
}