CATCH #503
DETECTED AUG 4, 2026 07:28 UTC
THE CATCHscore 10/10
EXHIBIT 503 · FLASH
Aug 4, 2026 07:28 UTC
web app
https://chatgpt.com
OpenAI's ChatGPT web app update reveals multiple new GPT-5 model IDs, including 'GPT-5-pro', 'GPT-5-reasoning', 'GPT-5.1', and 'GPT-5.2', alongside internal model picker features and experimental conversation endpoints.
READY TO POST:
looks like openai is prepping multiple gpt-5 variants! 'gpt-5-pro', 'gpt-5-reasoning-mini', 'gpt-5.1', and 'gpt-5.2' all spotted in a recent chatgpt web app update.
new chatgpt frontend strings hint at some serious internal development: 'gpt-5', 'gpt-5.1', 'gpt-5.2' model ids and 'internal models' search options are showing up. something big is brewing.
EVIDENCE — the receipts
⚠️ FROM WEB-APP BUNDLE — may include vendored-dependency strings, verify before posting.
ChatGPT web app (https://chatgpt.com) deployed new frontend code [bfe3fd5d37a71995 -> 601f4b0d0079c193]: +45943 strings, -1600 strings, 224 interesting
[model_id] (33)
+ /cdn/assets/gpt-details-gt8i8zk7.css
+ /cdn/assets/gpt-icon-h9k3y3p1.css
+ GPT-5
+ admin/model-access/gpt-5.6-preview
+ chatgpt.workspace.model.GPT-4.1.access
+ chatgpt.workspace.model.GPT-4.5.access
+ chatgpt.workspace.model.GPT-5-pro.access
+ chatgpt.workspace.model.GPT-5-reasoning-mini.access
+ chatgpt.workspace.model.GPT-5-reasoning.access
+ chatgpt.workspace.model.GPT-5.1-pro.access
+ chatgpt.workspace.model.GPT-5.1-reasoning.access
+ chatgpt.workspace.model.GPT-5.1.access
+ chatgpt.workspace.model.GPT-5.2-pro.access
+ chatgpt.workspace.model.GPT-5.2-reasoning.access
+ chatgpt.workspace.model.o3-pro.access
+ chatgpt.workspace.model.o4-mini-high.access
+ chatgpt.workspace.model.o4-mini.access
+ gpt-4l
+ gpt-4o
+ gpt-4o-gdrive
+ gpt-4o-gdrive-0212
+ gpt-4o-gdrive-0314
+ gpt-4o-gdrive-1007-mclick
+ gpt-4o-gdrive-1124
+ gpt-4o-gdrive-1124-citation
+ gpt-5
+ gpt-5-2
+ gpt-5-5-instant
+ gpt-5-pro
+ gpt-5-thinking
+ o3-mini-high-in-context-upsell-plus
+ routes/admin_.model-access.gpt-5[.]6-preview
+ ve reached the current usage cap for GPT-4. You can continue with the default model now, or try again {formattedTime}. <link>Learn more</link>`,values:{link:nH,formattedTime:a}}):(0,QV.jsx)(Q,{id:`TextMessageDisplay.gptUsageCapExceededCustomFeature`,defaultMessage:`You
[feature_flag] (102)
+ ))&&t.next_run_times.length===0?e.completed.automations.push(t):t.is_enabled?e.scheduled.automations.push(t):e.paused.automations.push(t),e),{scheduled:{label:xI.scheduled,filter:
+ ,action:E,actionMessageDescriptor:$L.focusPromptTextarea,group:Ka.Internal,bindings:[{binding:[Ko.Shift,Ko.Escape]}]},{key:
+ ,e[e.INTERNAL=1]=
+ /cdn/assets/contentful-dynamic
caught by the 20-minute sweep ·
the live wire · all receipts