gemini 3.1 pro openai dev day predictions:
* Persistent "o" Agents: Always-on AI that runs in the background.
* ChatGPT "Pro Max": A rumored $500/mo tier for extreme speed.
* "Ultrafast" APIs: Cerebras-powered processing at 750 tokens/sec.
Opus 5.5 - top 5 OpenAI DevDay predictions:
* "o", an always-on agent
* Ultrafast API tier
* Managed agents + multi-agent Codex
* $500/mo ChatGPT Pro Max
* ChatGPT + Work merged into one app
@ChatGPT astra for hard surface modeling, took 18min:
* imagegen 2.5 with gpt to create concept art, then
* codex astra high with blender-cli skill for @Blender modeling