MANY MINDEDOpenAI slashes GPT-5.6 developer pricing under Chinese competition
OpenAI has cut developer pricing across its GPT-5.6 line by as much as 80%. Sam Altman announced the change on X on Thursday: the lightweight Luna model falls 80%, to US$0.20 per million input tokens and US$1.20 per million output tokens. The mid-tier Terra model drops 20%, to $2 and $12 per million input and output tokens. The flagship Sol gains a fast mode that OpenAI says raises performance speed up to 2.5 times. All three models were introduced only earlier this month.
The cut moved Luna into the most attractive tier of Artificial Analysis's intelligence-per-dollar ranking, above Chinese rivals including Zhipu AI's GLM-5.2 and MiniMax's M3 — which is the whole point of it. SCMP reads the move as defensive: open-weight and cheaper Chinese models have been winning international developers, and OpenAI answered on price rather than on capability.
Competitive price cuts are the cleanest signal this project tracks, because nobody chose them. The price of machine cognition fell because a rival made it fall. An 80% cut at the low end is roughly the difference between metering API calls and not thinking about them, which is the point at which a capability starts behaving like infrastructure rather than a purchase.
The caveat is that this is list pricing on a hosted API, not a cost of production, and list prices can move back up. Watch whether the Chinese labs answer, and whether these rates hold past this round.
Source: SCMP Tech