Grok 4.5 has just become a lot cheaper without anybody noticing or publishing anything, and it’s now the best “balanced” model to use for agentic tasks according to our internal benchmarks.
Grok launched a few days ago with a price for input/cache/output of 2$ / 0,50$ / 6$. Looks cheap, but there’s a catch: the cache is very expensive, and it’s what makes up to 80-90% of used tokens in a real world agentic task. For reference, GPT Terra high’s pricing is 2,50$/ 0,25$ / 15$. Despite being apparently a much more expensive model, it's actually cheaper in real world tasks.
However, when I run our Vibetasking benchmark tests today, Grok was the cheapest model (even cheaper than GPT 5.6 Luna).
Trying to understand why, I found out Grok 4.5 cache price is now 0,30$ and I see no trace on the internet of anybody saying anything about it.
With this, Grok is a lot cheaper, while being extremely smart, becoming the best model by far to use as the reasonable, balanced option for most tasks, and really is on the Pareto Frontier now.
verdverm•59m ago
> without anybody noticing ... I see no trace on the internet of anybody saying anything about it
There are well known reasons most people ignore Grok, regardless of any technical merits it may or may not have.
timoteimolnar•32m ago
What are some of these reasons? Previous models have been shit, but this one's solid
verdverm•13m ago
The non-technical reasons, like the person behind it, the bias injected into the model, and it's usage to create non-consensual pornography of real people (and more so the attempted legitimization to this usage by person behind it)
There are plenty of solid models, many of us have no interest in using the one proffered by the world's richest megalomaniac
timoteimolnar•1h ago
Grok launched a few days ago with a price for input/cache/output of 2$ / 0,50$ / 6$. Looks cheap, but there’s a catch: the cache is very expensive, and it’s what makes up to 80-90% of used tokens in a real world agentic task. For reference, GPT Terra high’s pricing is 2,50$/ 0,25$ / 15$. Despite being apparently a much more expensive model, it's actually cheaper in real world tasks.
However, when I run our Vibetasking benchmark tests today, Grok was the cheapest model (even cheaper than GPT 5.6 Luna).
Trying to understand why, I found out Grok 4.5 cache price is now 0,30$ and I see no trace on the internet of anybody saying anything about it.
With this, Grok is a lot cheaper, while being extremely smart, becoming the best model by far to use as the reasonable, balanced option for most tasks, and really is on the Pareto Frontier now.
verdverm•59m ago
There are well known reasons most people ignore Grok, regardless of any technical merits it may or may not have.
timoteimolnar•32m ago
verdverm•13m ago
There are plenty of solid models, many of us have no interest in using the one proffered by the world's richest megalomaniac