Why was o-1 and o-1 mini models deprecated when it was working perfectly well ?
Do they have to deprecate one when they introduce new ones ?
I can understand if they need to “upgrade ChatGPT” with the latest model of their choice.
But why cant o-1 mini be accessible via APIs ?
Perplexed user Dev
Since you are not likely to get an answer directly from a decision-maker…
As we launch safer and more capable models, we regularly retire older models.
o1 (o1-2024-12-17) is not deprecated, only o1-preview at this point.
o1-preview and o1-mini are notable for lacking Responses endpoint support, or alternately missing features like image input, function-calling and structured outputs, etc.
gpt-4.5-preview, “we need those GPUs back, guys…”
They don’t have to deprecate…
Deprecation would imply there is planning time to allow migration and adaptation. Instead, older reasoning models were instantly replaced.
Possible motivation there - this “speed” ~= “computation cost”:
Ok, thanks.
I kinda looked at these comparisons before. Know about the light bulbs and lightning bolt comparisons.
