Must say I’m very impressed with the Grok 4.5 High Fast model.
Listens well to prompts and seems to read all rules, love it so far!
Alright Grok 4.5 is now live in Grok Build for EU.. at what point are we getting it for Cursor?? im in EU..
Grok 4.5 is stupid-good for the price. I’d say it’s basically at Opus 4.8 level and has far superior tool calling abilities. Please note: I am only judging this based on planning / generating / documenting code for - what I would consider to be - relatively simple applications. I’m sure Opus 4.8 (and Fable 5, of course) are far superior in terms of overall knowledge, but Grok 4.5 is the best model I’ve ever used for coding (price/performance).
AGAIN PLEASE NOTE: your mileage may vary. I have a very specific workflow that is designed to minimize AI code slop. If my prompting was more freeform, I suspect quality of output would drop. For my use-cases, though, Grok 4.5 is an absolute beast.
Will we get 50% off in the EU too? It’s available only as of today and the discount period ended yesterday..
I’m curious when Auto and Composer 2.5 are updated, how integrated Grok will be?
My preference would be to not use Grok. I am hoping there will be Grok-free first-party model options.
If you don’t want to use the better of the two available models, you can just stick with Composer 2.5
1.) I’ve used Grok 4.5 for GEANT4 research and simulations on antimatter propulsion, custom AI training, and Grok 4.5 taught me to solder!
2.) In my experience Grok handles long-running complex tasks very well! Grok managed, iterated, and ran thousands of simulations for me while I slept!
3.) Grok is insanely fast. Not only fast, but accurate and efficient.
Pretty impressive results so far with Grok 4.5. I assume it’s considered 1st Party model like Composer. I do like Grok switching automatically to plan mode for more complex tasks where Composer just starts building (and usully it builds ■■■■ then…). In normal day 2 day ops we run most on Opus so far but Grok seems to be a real alternative for it
Hey all!
A couple of updates.
- Grok 4.5 is now available in the EU!
- We’ve extended promotional pricing for Grok 4.5 (50% off, or double usage) through July 21nd
- We’ve also doubled the included usage of Cursor models on all plans!
Great, and again the legacy plan users can not use it without MAX mode 
I see that it is fast, but I’m not sure why I would choose a certain model unless I fail at doing something with auto mode. I feel auto mode should be able to determine which model to use, so I’m confused why would I specifically use a model unless I know that this problem can only be resolved by that particular model, but there is no data which exist to help me with that.
Did anybody test Grok 4.5 vs Kimi K3 for real ? The current benchmarks and tests I saw on YT do not favor Grok for sure …
@Colin the double included usage that’s like always from now on? Not a promo? And how is this done during the current month percentage counts, since I didn’t see it becoming less percentage in my spending?
In terms of cost, no, considering the quota size we get here on Grok and Composer. And according to the metrics, Grok should be faster.
It’s not a promo, and the increased usage already applies to this month!
In panning mode Grok 4.5 doesn’t trigger then proper question UI and just asks via text