Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Cursor Introduces Composer 2.5

cursor.com

Post reply on HN
Print view View original

201–210 of 238 posts

Re: Cursor Introduces Composer 2.5

#201

Post by jmcqk6 »

onlyrealcuzzo wrote: post #23
asar wrote: post #2

The model is (like Composer 2) based on Kimi K2.5 and they claim SOTA performance for 1/10th of the cost. The tweet also mentions that they've started a new model from scratch on Colossus 2 (xAI/SpaceX Cluster). Really impressive how they've made this jump from being called the vscode fork with no moat just a couple of months ago.

> Really impressive how they've made this jump from being called the vscode fork with no moat just a couple of months ago. Impressive, yes. But they still don't have a moat...

I've been using cursor for over a year for my personal projects. At work, I use Claude Code, and so I've been wondering if I'm missing something in the other agents.

Over the last week, I tried out two other agents on my personal projects: dirac and forgecode, after seeing impressive results from both of them on terminal bench.

After a good amount of testing, and over $100 in open router spend, I'm back to cursor.

I really liked forgecode the best, and it feels better than claude code, but cursor definitely feels best to me. Composer 2.5 is fast and effective, and it makes a huge difference. I was running `forge` with Opus, and it was taking dozens of minutes to do things, and the feedback loop was so slow.

The previous version of composer was also much faster, and it makes a difference. Maybe people like context switching, but I prefer to stay focussed on the task in front of me, and I'm reviewing the code carefully.

I think that's a pretty good moat. I was ready to end my subscription a week ago, and now I'm back after learning the grass is not necessarily greener on the other side of the fence.

Re: Cursor Introduces Composer 2.5

#202

Post by SebastianKra »

whywhywhywhy wrote: post #104
goyozi wrote: post #62

I kind of want to try it, to see if and how far they can take an open model and improve it but I really don’t miss the Cursor user experience. Constant UI changes, half-baked features, smaller and smaller limits, useless AI change attribution; I think I’ll wait for others to report if it’s any good.

Noticed recently they keep opening their “Agents” window when the project was last opened in the VSCode fork window in the hopes I’ll just continue working in that when the UI is totally different and missing things I need. For a professional tool it’s getting egregious how little respect they have for my workflows and flow state they way they keep moving, changing iconography and flipping switches of the UI. It’s cl…

It seems obvious that they plan to eventually drop VSCode. I'd be willing to take them up on that offer. Their agent window is genuinely better as a starting point.

What annoys me is how little they want to integrate with ...anything. Wanna open a link in your default browser? Use our built-in chromium fork, we insist. Wanna open a location in Zed? No, please use our half-baked editor re-implementation. Wanna open a location in Cursors own vscode-based editor? You can't. Managed to work around that somehow? We changed your files to "Worktree TS", disabling all your language servers. It's like programming on an iPhone.

Re: Cursor Introduces Composer 2.5

#203

Post by jmcqk6 »

PUSH_AX wrote: post #8

They set themselves up for flack when they use whatever these evals are… they did the same for composer 2 which was evaled in close competition with frontier models, spoiler alert, it wasn’t even close in practice. So now 2.5 is supposed to compete with opus 4.7? Sure…

That does not match my experience. Composer 2 was fantastic for my uses, and I hit Composer 2.5 with some very difficult things last night, which it handled fast and effectively. I don't really care about benchmarks. I care about practice, and in practice, it's been very very good for me.

Re: Cursor Introduces Composer 2.5

#204

Post by dewey »

varun_ch wrote: post #161
wesammikhail wrote: post #160

Earlier quoted context omitted.

I dont know what it is with products names these days. Antigravity, Antimatter, Composer, Clay, Ramp, Bolt, etc. You'd think the founders would Google for naming conflict before choosing a name.

I genuinely wonder if consulting LLMs for naming advice could be an explanation. They certainly wouldn’t be great at coming up with new words for a product name.

Naming issues are as old as time. Apple Computer vs. Apple Records comes to mind as a popular example.

Re: Cursor Introduces Composer 2.5

#205

Post by wunderlotus »

I love Cursor as a tool, but I'm skeptical bc:

1/ CursorBench is so opaque [1] that it makes it hard to trust. Not to mention the v3.1 eval is a newer iteration and there's no insight into the tasks or if the model was just tuned to max it out. Composer 2 previously scored between 60-65% on the previous benchmark eval [2] but scores between 50-55% on CB v3.1[3].

2/ I've experienced Composer 2's performance and it leaves much to be desired as a daily driver for a knowledge worker. but KWs are obviously not the target users and I can see how it's cost-efficient for executing on clearly-defined, discrete coding tasks. Obviously that's their value proposition and they're figuring out how to communicate it well to the target customer. It just doesn't feel like CursorBench is that.

[1] https://cursor.com/blog/cursorbench#building-cursorbench

[2] https://cursor.com/blog/composer-2-technical-report#performa...

[3] https://cursor.com/blog/composer-2-5

Re: Cursor Introduces Composer 2.5

#206

Post by chemex »

I've been using Claude Code as my daily driver on a React Native + iOS codebase for the last few months. The thing that surprised me wasn't quality differences on individual edits — those are pretty close once you control for harness wiring — but how differently I'd ended up structuring my workflow around each style of tool.

Tab completion + chat-in-sidebar feels like an extension of my editing. An agentic harness feels more like delegating a 20-minute task and coming back to review. Different cognitive load, different bug profile. The "which is better" framing tends to skip over the fact that they reward different working styles.

Two things I'd watch on Composer 2.5 specifically:

1. How it handles long-running multi-file refactors that touch 10+ files. My experience with smaller models in that slot is they lose track of which files they've already edited around 30% of the way through. Frontier models keep the plan coherent for longer.

2. How it deals with non-obvious file boundaries. The thing that takes me out of "let it work" mode is the model deciding it needs to edit a config file I didn't think of. Usually that's right, but occasionally it's spelunking somewhere I don't want it to be.

The Kimi K2.5 base is interesting on its own. Open weights below frontier closed models is the thing worth watching from the harness side. If anyone's set up to fine-tune for a specific harness, this is the moment.

Re: Cursor Introduces Composer 2.5

#207

Post by fjdjshsh »

goyozi wrote: post #62

I kind of want to try it, to see if and how far they can take an open model and improve it but I really don’t miss the Cursor user experience. Constant UI changes, half-baked features, smaller and smaller limits, useless AI change attribution; I think I’ll wait for others to report if it’s any good.

I've had good experiences with Cursor so far and it's my main IDE. I've noticed some UI changes, but I've switched fast and they didn't bug me

Re: Cursor Introduces Composer 2.5

#208

Post by intrasight »

mwigdahl wrote: post #142
intrasight wrote: post #139

Earlier quoted context omitted.

Is that true? If the distillation is not lossy and the model runs much faster due to less resource consumption, then it may outperform.

One of those conditionals is a pretty huge assumption.

It's an assumption and it can be tested

Re: Cursor Introduces Composer 2.5

#209

Post by chis »

chemex wrote: post #206

I've been using Claude Code as my daily driver on a React Native + iOS codebase for the last few months. The thing that surprised me wasn't quality differences on individual edits — those are pretty close once you control for harness wiring — but how differently I'd ended up structuring my workflow around each style of tool. Tab completion + chat-in-sidebar feels like an extension of my editing. An agentic harness fe…

AI slop detected, you're under arrest

Re: Cursor Introduces Composer 2.5

#210

Post by WhitneyLand »

pdq wrote: post #178
WhitneyLand wrote: post #155

Say what you want about Cursor but they don’t lack for ambition. Forking VS Code, going big on bleeding edge features like cloud agents, and now they’ve thrown down the gauntlet directly challenging frontier labs by training their own model (“much larger” than Kimi 2.5’s 1T parameters) from scratch. They’ve been highly successful so far. Raised $50B, $2B in revenue, forecast to end 2026 above $6B. But even at these h…

Why is this comment upvoted? It is most likely AI generated with a nice "Raised $50B" hallucination and filled with cliches ("thrown down the gauntlet", "mountain you don’t climb just once", "not for the faint of heart").

I wrote this 100% off the top of my head on my phone while eating a sandwich.

Ffs.

edit: removed cursing you out. Sorry but this is frustrating. I don’t leave AI generated comments here (or anywhere else).

Post reply on HN