Ddos 
Yesterday, OpenAI announced a dramatic 80% price reduction for its o3 reasoning model, lowering the cost per million input tokens from $10 to $2, and per million output tokens from $40 to $8 when accessed via API.
This significant price cut makes the o3 model markedly more appealing to developers, especially given its performance advantages over many competing models. Historically, the primary drawback of OpenAI’s offerings has been their comparatively high cost.
Following the price adjustment, concerns emerged among developers that the reasoning capabilities of the o3 model may have been silently downgraded. In response, OpenAI issued a statement to reassure the community: “We optimized our inference stack that serves o3. Same exact model—just cheaper.”
While most general users do not interact with these models via the API platform, pricing is a critical factor for developers. Tools that rely on API integration, such as Cursor and Windsurf, stand to benefit from a substantial reduction in operating costs.
Skepticism lingers among some developers, but independent benchmark testing by the ARC Prize community has confirmed that the performance of the o3 model remains consistent with its original iteration. This affirms that OpenAI has not substituted the model to facilitate the price drop.
Additionally, OpenAI unveiled a new flagship reasoning model yesterday—o3-pro—currently available exclusively to ChatGPT Pro and ChatGPT Team subscribers. Although accessible via API, its high pricing renders it a less cost-effective option for broader use at this stage.
Donate