I think they should add GLM 4.5, Qwen-3 coder and Kimi k2.
1 Like
Hi @deanrie, this request currently ranks 4th among all feature requests now. Do you think it has enough votes to move forward?
3 Likes
Up up up up
I think yes. Thanks everyone for your interest in this model. I’ll check with the team to see if we can add it.
11 Likes
What’s the problem with highlighting alternatives?
Actually no problem, I just don’t think these models are worth much effort to integrate into Cursor while Gemini be superior than them.
yeah well gemini does not work
these models have a potential, if cursor adds one of these new models on one of those fast api, you can generate whole projects in seconds for couple of cents at most, see if it can handle it and if not, then switch to gemini
win win situation
I really hope that the release of gpt 5 won’t cancel this feature, and we will see GLM 4.5 / Qwen3 coder 480b in cursor soon… (and I’m not talking about auto mode, just wanna use them in cursor without having to break all openAI models by using openrouter endpoint)
My Cursor Pro subscription was only for 3 days? Comedy. GLM 4.5 needs to come out urgently. No one cares about ChatGPT 5 because that’s going to expire in 3 ridiculous days.
hi i have 2 questions.
1. i saw there was many votes to add glm 4.5 to cursor as auto mode since the auto mode was really not so good… and i voted too for it so wanted ask happened to progress?
- after watching much about GLM 4.5 and its capabilities how good it is, i starting to wondering what is the justification to not add it already?
i mean if its so cheap, open source, small and super quality, then why is it still not added immediatly? i think already that the quality of claude opus / price is almost a scam compared to other models i thought in my experience when i used up the credits in 2 days just and was around same quality like sonnet that required to rerun many times to fix the code…. well i dont understand what is taking so long. i just really wish to have a flatrate option to use for vibe coding… and not worry all the time to waste credits and not succeed… but 200 is just unaffordable to me, even 60 is too much and i really thought that when glm is so good then finaly will be able to have real coding flatrate as vibe coder. 
shouldnt the glm 4.5 and qwen coder now be the only options as auto mode?
2 Likes
@condor any inside info on this? we are missing a lot of good and efficient models
hi @weeion_photoniac @liquefy we were really busy with already planned model and feature launches as you can see there is now Cursor 1.4, cursor-agent (CLI) and GPT-5 was added.
As my colleagues mentioned we are looking at new models regularly and they need to be evaluated. While I have not tried GLM 4.5 some models, while working well in API or chat might not offer good tool support etc. Therefore before using them in Auto we need to actually make sure that we are not lowering quality of Auto with new additions.
Please bear with us while we are working hard to bring you more models and better Auto.
4 Likes
any update about GLM 4.5 models ? why still not being added ? tired of waiting. this model is so good close enough to sonnet 4.
2 Likes
hi, any update about GLM 4.5 model ?
1 Like
Hey, thanks for the request. We’re considering it, but the timing for adding this model is still unknown.
1 Like
Now you know why they delayed adding this model to auto mode.
please add GLM 4.5
1 Like
why ?
New auto pricing. People would go mad if glm 4.5 is at 1.25/6$ pricing.
@wavification That’s not quite accurate. We’re always open to adding well-performing models. However, any new model must demonstrate strong performance within Agent, as we’re committed to maintaining high output quality, even in Auto mode. Thorough testing is required, which means coordinating with providers for any necessary adjustments, making changes on our end, and ensuring sufficient capacity. This process does take time, but it’s crucial for upholding our standards.
2 Likes