As I was reviewing my automations, I stumbled upon a shocking discovery: some legacy models in the API have increased in price by a whopping 100x. Yes, you read that right – 100x! This got me thinking, and I wanted to share my experience to spare others from a similar surprise.
The culprit was the `gpt-4-1106-preview` model, which now costs a staggering `$10/M input tokens + $30/M output tokens`. Meanwhile, the GPT-4o Mini model costs a relatively affordable `$0.15/$0.60`. I was using the legacy model for mailbox analysis and lead processing, which processed a significant number of emails daily. Since the price was low initially, I was comfortable passing large context at a time. But now, I’m glad I caught this before it burned a hole in my pocket.
The lesson learned is to regularly review your scripts and pricing pages to avoid such costly surprises. It’s easy to overlook these changes, but it’s essential to stay vigilant to avoid financial shocks.
So, take a moment to double-check your API usage and pricing. You never know when a legacy model might be hiding a hefty price tag.