4 ms·
They say it's twice as fast/cheap, which might matter for your use case.
by whyever 2y ago
They say it's twice as fast/cheap, which might matter for your use case.
- minimaxir 2y agoIt's twice as fast/cheap relative to GPT-4-turbo, which is still expensive compared to GPT-3.5-turbo and Claude Haiku. https://openai.com/api/pricing/ https://openai.com/api/pricing/
- c0t300 2y agobut better afaik
- minimaxir 2y agoBut may not be better enough to warrant the cost difference. LLM cost econonmics are complicated.
- BeetleB 2y agoFor commercial use at scale, of course cost matters. For the average Joe programmer like me, GPT4 is already "dirt cheap". My typical monthly bill is $0-3 using it as much as I like. The one time it was high was when I had it take 90+ hours of Youtube video transcripts, and had it summarize each video according to the format I wanted. It produced about 250 pages of output. That month I paid $12-13. Well worth it, given the quality of the output. And now it'll be less than $7. For the average Joe, it's not expensive. Fast food is.
- fvdessen 2y agoI’d much rather have it be slower, more expensive, but smarter
- pests 2y agoThen the current offering should suffice, right?
- specproc 2y agoDepends what you want it for. I'm still holding out for a decent enough open model, Llama 3 is tantalisingly close, but inference speed and cost are serious bottlenecks for any corpus-based use case.
- abdullin 2y agoI think, that might come with the next GPT version. OpenAI seems to build in cycles. First they focus on capabilities, then they work on driving the price down (occasionally at some quality degradation)