Feature request for product/service
AI Models
Describe the request
We’re missing out on huge value, we’re currently paying a lot more for the same intelligence.
+1 
Guys try to like and comment more on this post so this may attract attention and may get implemented
pls add both GLM-5.3 and GLM-5.3 Flash!
please!
+1, but I know you guys are trying to run a business here. At an absolute minimum, please release Composer 3 built on top of GLM-5.3 Flash. I don’t care if it literally has zero Cursor-specific optimizations and it’s just the exact same model with a Composer sticker on top of it. Just please do it as soon as possible.
I am already using it with BYOK, but it kills built in models 
+1
Please add this. Remove GLM 5.2 to make space for it 
I need this yesterday!
Actually guys upvote the post itself instead of emoji likes, upvotes push a post to the top for the cursor team
The story with third-party models will end soon. You know you won’t get GLM 5.3 Flash.
All you will have soon is Grok Lite, Grok Pro and Grok Ultra. Cursor has investors.
- If you have nothing good to say then abstain from saying it, we’re trying to show the team that there is support behind this feature request, and the post is slowly growing.
- Even if that’s true, if cursor chooses to go exclusive only with Grok models then competition will rise, people will flock for other AI IDE’s, Cursor will have to play a balancing act.
I think adding the open models would be a good response to OpenAI removing their models. You should become to best Grok+Open Models Harness out there.
the value you could get of 5.3 flash is insane, PLEASE add
+1 +1
one more plus! add it pls!
Does anyone know why the Cursor team doesn’t add really good open-weight models? They haven’t added the new DeepSeek V4 and V4 Flash models yet.
GLM-5.3 Flash is much better than GPT-5.6 Luna, and since the OpenAI team has already announced that they would stop offering their models through Cursor, this could be one of the best replacements.
This is speculation on my part, but I both strongly believe what I’m about to say and actually don’t blame them since I understand it to some extent: It’s because these models significantly impact their bottom line.
Besides subscriptions, Anysphere likely makes most of its money through margins on the token usage of customers. They almost certainly get discounted rates for usage through providers like OpenAI and Anthropic, and thus the difference in that pricing compared to what users get charged likely makes up a lot of their revenue (or technically cost-savings, since it’s not exactly income).
Bringing that back to these open-weight models, not only do these tend to be cheap as it is - since open-weight pricing tends to be that way - they likely also can’t get much of a discounted rate on them, thus lowering their margins further.
This is the most speculative part of my reply so far, but I think back to when everyone was begging for Kimi K2 and Kimi K2.5 at the time. If I recall correctly, neither of those were really made available in Cursor until the relevant Composer models were released to be able to compete with them. If you look at the open weight models we DO have right now, they’re either not cheap (Kimi K3) or are undercut by Composer 2.5.
That’s why even in this thread, I personally don’t actually ask for GLM-5.3 Flash - I asked for Composer 3.
+1 glm 5.3 / flash need
You know what, at this point am fine with composer 3 as long as it performs like GLM-5.3 Flash and is relatively at the same cost, i don’t care what models we use but as for now the models provided by Cursor have become outdated with the release of GLM-5.3 Flash.
All am asking for is for us, the developers, to remain at State Of The Art LLM models via Cursor supplying them to us.
If Cursor fails to supply that, chances are devs will grow frustrated and look somewhere else. Time’s ticking!