Even without an imminent release it's a good strategy. They're getting pressure from Qwen and other high performing open-weight models. without a horse in the race they could fall behind in an entire segment.
There's future opportunity in licensing, tech support, agents, or even simply to dominate and eliminate. Not to mention brand awareness, If you like these you might be more likely to approach their brand for larger models.
Is this the stealth models horizon alpha and beta? I was generally impressed with them(although I really only used it in chats rather than any code tasks). In terms of chat I increasingly see very little difference between the current SOTA closed models and their open weight counterparts.
> I can't think of any reason they'd release this unless they were about to announce something which totally eclipses it
Given it's only around 5 billion active params it shouldn't be a competitor to o3 or any of the other SOTA models, given the top Deepseek and Qwen models have around 30 billion active params. Unless OpenAI somehow found a way to make a model with 5 billion active params perform as well as one with 4-8 times more.
From a strategic perspective, I can't think of any reason they'd release this unless they were about to announce something which totally eclipses it?