GLM 5.3 and GLM 5.3 Flash are now available in Cursor! GLM 5.3 Max is the best-scoring open-weight model on CursorBench 4.0.

Oct 1, 2026 · 10:30 PM UTC

201
171
3,585
234,509
Sort replies: Relevant Recent Liked
Replying to @cursor_ai
When can I access Cursor with Supergrok?
569
Replying to @cursor_ai
Open weights eating the ladder.
1,251
Replying to @cursor_ai
@grok is that part of the cursor models in the subscription? or other models?
1
6,686
Replying to @cursor_ai
Better than 5.6 Sol and nearly Fable 5.1 Low level? Like cmon it's a good model but not that good.
3
472
Replying to @cursor_ai
oh yes let me pay for glm 5.3 when its free om nvidia 😩😩 les go
1
3
3,211
Replying to @cursor_ai
So now only Grok 4.6 medium for Free users? I love𝗱 you Cursor (yea, love𝗱)
1,086
Replying to @cursor_ai
@grok are these hosted on cursor/xai server? Or from Chinese API?
1
1
1,760
Replying to @cursor_ai
This is actually disgusting. Glm 5.3 flash is 50% off 0.25 on openrouter. It is far worse than Luna at 0.50. At this point, might as well support openrouter models to get access to all openai gpt-6.1. It is not like you don't take a cut already.
1
524
Replying to @cursor_ai
Stop with the backwards X axis. Please stop.
3
77
3,503
Replying to @cursor_ai
Now we balling!
961
Replying to @cursor_ai
I think it’s a big win for subscribers.
2,865
Replying to @cursor_ai
So GLM 5.3 will be using which usage bucket?
1
1,873
Replying to @cursor_ai
open weight topping cursorbench is wild
1
1,118
Replying to @cursor_ai
GLM 5.3 flash is my fave for quick & dirty coding line by line
1
1
2,239
Replying to @cursor_ai
Kimi k3 not even on the chart?
2
15
3,279
Replying to @cursor_ai
I mainly use Auto, composer, and Grok 4.7
2
2,485
Replying to @cursor_ai
that is huge news!! I was thinking man it'd be awesome if they add some cheap intelligent openweight models
2
3,465
Replying to @cursor_ai
Now show GLM flash and also include composer 2.5. Composer 3.0 when??
2
1,276
Replying to @cursor_ai
Still far too expensive. Grok 4.5 and composer 2.5 is where i will stay.
1
4
1,343
Replying to @cursor_ai
Any plans to test Astra 6.0 and Sol 6.1?
1
1,692
Replying to @cursor_ai
GLM 5.3 is not good. Try Qwen 3.8 Max
1
142
Replying to @cursor_ai
中文版对比：
耳朵
@RookieRicardoR
4h
GLM 5.3 上架 Cursor，是 CursorBench 评分榜上最强的开源模型，价格大概是 Claude Sonnet 5.5 的一半， 对比一下 Cursor 中的其他开源模型，Kimi K3 目前的价格是 Claude Sonnet 5.5 的 1.5 倍，GLM 还是很有性价比的。
182
Replying to @cursor_ai
Why did it take so long 😅
3
923
Replying to @cursor_ai
Doesn't seem to be much of a point compared to Grok 4.7
17
Replying to @cursor_ai
open weights, closed wallet
1,770
Replying to @cursor_ai
Are you competing for getting into the most expensive and less performant section? 🤦‍♂️ And Flash is your only competitor?
1
2,751
Replying to @cursor_ai
is it actually cheaper than sonnet though
5
1,951
Replying to @cursor_ai
Why waste my other models pool on anything else other than Opus 5.5? It's too small.
1
215
Replying to @cursor_ai
25 cent penalty in GLM 5.3 Flash charts though :(
1
420
Replying to @cursor_ai
Got it: Sonnet 5.5 High + Opus 5.5 Med/High
94
Replying to @cursor_ai @tibor_tee
How is the token usage?
1
1
424
Replying to @cursor_ai
Where’s 5.3 flash on the plot?
10
1,382
Replying to @cursor_ai
@Cursor - are you working on a Composer 3.0? I'm quite happy with the 2.5 - but a few things could be improved.
6
689
Replying to @cursor_ai
I don't believe that Sonnet can perform better than Fable...
887
Replying to @cursor_ai
Anthropic said how much revenue came from their top two customers? Look at the opus/sonnet lead my god
152
Replying to @cursor_ai
At this point, this should be a daily update.
552
Load more