Apparently better and way cheaper than Sonnet? To be seen…
Yes. please add this model !!!
i wonder how DeepSeek R1 would perform with user grade laptops under cursor
yes,DeepSeek is cheaper and better than Claude Sonnet 3.5.
but the question is:
Deep Seek 's context token is longer than Claude Sonnet 3.5?
that the true key.
yes please, please, please get us this model inside Cursor
I had an interesting daydream just yesterday that maps to your comment:
“AI 'Droplets” –
An AI Agent deployed to an IoT device that when empowered had enough agency and connetivity to be an AI IoT node that is a more sentient sensor
(Disclaimer: I was Dir IT OPs for Big DefenseCo that built stuff like this)
–
Now we have talked about Edge Compute forever, but Edge reasoning is novel.
- SO even if we can micro_focus the bot via an even more distilled R1_planck version which has cellular automata logic on GOTO
00 Im alive
01 I have a persona
02 I have skills
03 I have data
04 I have a goal
05 I have connections
06 I have a rules and direcctives
07 I have a task
08 I have a reporting stream
09 I have to provide status
10 GOTO 00
===
Awareness is mycelial in nature -
I love emergent behavior and growth.
can we have it team @cursor ??
DeepSeek v3 is now available to test in the editor on the latest version of Cursor.
You will have to go in and enable it in your model settings first, but it is included as a non-premium model, so give it a try.
Please note: this is an experimental model, so please note it could be removed at any time.
does it not support composer or is it the official version that supports it?
It’s really great!!! When will it support R1?
there’s a very easy way for cursor to handle this
It works with normal
composer. agent
composer is currently only supported using Anthropic models or GPT-4o.
can normal composer do yolo mode?
No, because normal composer can not access the command line. So there is nothing to yolo
( )
I know that but I don’t know if it will support agent when it is official version
I would imagine that it depends on how well users receive this new model, balanced with the cost.
if it is officially released then there will be less claude users and people will stop complaining about slow requests lol
Have been testing R1 all afternoon, the model is amazing at a fraction of o1’s price. Please add.
following, i would love to try it out, seems an amazing model also so cheap to use