Two interesting GPT-5.6 Sol changes just happened back to back:
→ OpenAI changed its positioning from “Latest frontier agentic coding model” to “Reliable agentic workhorse for everyday tasks,” while moving its catalog priority from 6 → 4.
→ It also removed the “Ultrafast” service tier from GPT-5.6 Sol, even though Codex still keeps explicit Ultrafast support internally.
My current assumption is that OpenAI is clearing space in the model lineup for GPT-6 Sol, and may be preparing to make Ultrafast a service tier specifically for GPT-6 Sol instead.
If you missed the context, I documented both changes with the exact GitHub diffs in my previous two posts.
显示更多