frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: Sweep, Open-weights 1.5B model for next-edit autocomplete

https://huggingface.co/sweepai/sweep-next-edit-1.5B
158•williamzeng0•7h ago
Hey HN, we trained and open-sourced a 1.5B model that predicts your next edits, similar to Cursor. You can download the weights here (https://huggingface.co/sweepai/sweep-next-edit-1.5b) or try it in our JetBrains plugin (https://plugins.jetbrains.com/plugin/26860-sweep-ai-autocomp...).

Next-edit autocomplete differs from standard autocomplete by using your recent edits as context when predicting completions. The model is small enough to run locally while outperforming models 4x its size on both speed and accuracy.

We tested against Mercury (Inception), Zeta (Zed), and Instinct (Continue) across five benchmarks: next-edit above/below cursor, tab-to-jump for distant changes, standard FIM, and noisiness. We found exact-match accuracy correlates best with real usability because code is fairly precise and the solution space is small.

Prompt format turned out to matter more than we expected. We ran a genetic algorithm over 30+ diff formats and found simple `original`/`updated` blocks beat unified diffs. The verbose format is just easier for smaller models to understand.

Training was SFT on ~100k examples from permissively-licensed repos (4hrs on 8xH100), then RL for 2000 steps with tree-sitter parse checking and size regularization. The RL step fixes edge cases SFT can’t like, generating code that doesn’t parse or overly verbose outputs.

We're open-sourcing the weights so the community can build fast, privacy-preserving autocomplete for any editor. If you're building for VSCode, Neovim, or something else, we'd love to see what you make with it!

Comments

plutodev•3h ago
Nice work the next-edit framing matches how real refactors happen much better than token-level autocomplete.

The diff-format insight is especially interesting. Smaller models struggling with unified diffs lines up with what I’ve seen too simpler original/updated blocks reduce noise and improve intent capture.

On the infra side, training a 1.5B model in ~4 hours on 8×H100 is impressive. For folks experimenting with similar mid-scale models, we’ve been running comparable workloads on decentralized GPU aggregators (I’ve used io.net) to avoid cloud quota limits and keep costs predictable with the tradeoff that you handle orchestration yourself.

Curious if you saw diminishing returns when including older edits as context? That cutoff seems tricky in larger repos.

kouteiheika•34m ago
> On the infra side, training a 1.5B model in ~4 hours on 8×H100 is impressive.

It's hard to compare without more details about the training process and the dataset, but, is it? Genuine question, because I had the opposite impression. Like, for example, recently I did a full finetuning run on a 3B model chewing through a 146k entry dataset (with 116k entries having reasoning traces, so they're not short) in 7 hours on a single RTX 6000.

kamranjon•2h ago
I read the release but didn't quite understand the difference between a next-edit model and a FIM model - does anyone have a clear explanation of when to use one over the other? I'd love if there was a sublime plugin to utilize this model and try it out, might see if I can figure that out.
mgz•2h ago
I use Sweep’s Jetbrains autocomplete plugin daily, it really stands out.
8n4vidtmkvmk•1h ago
Better than the one that ships with Jetbrains?

I did buy their $100/yr AI but its about to run out.

dcreater•2h ago
Based on qwen2.5-coder? seems like a "why not/resume embellish/show VC" type release I guess
dang•2h ago
"Please don't post shallow dismissals, especially of other people's work. A good critical comment teaches us something."

https://news.ycombinator.com/newsguidelines.html

ing33k•2h ago
can it be integrated in monaco editor ?
bangaladore•2h ago
So SFT cost less only low hundreds of dollars? (1-10$ per hour per H100 if I'm seeing this correctly).

What about SFT?

Presumably basing this of Qwen is the reason it can be done for so cheap?

syntaxing•2h ago
Wow super fun read, I love how it went into the technical details. Any way to make it work with vscode?
martianlantern•1h ago
This is cool! I am more interested in how you guys generated next edit training data from repos, seems like there are lots of caveats here. Would love your insights

Again amazing work! waiting for what you guys cook next

sim04ful•1h ago
I'm very green to this so forgive if this question sounds silly:

Would instead of the RL step a constrained decoding say via something like xgrammar fix syntax generation issue ?

NitpickLawyer•1h ago
> Would instead of the RL step a constrained decoding say via something like xgrammar fix syntax generation issue ?

It can, but you have to consider two things here:

a) constrained decoding ensures adherence to syntax, not semantics. Say you're editing a field in an enum in rust. You can write syntactically correct rust code that doesn't address the new field further in the code (say in a switch). You'd get correctly syntactic code, but the compiler will scream at you. RL works on both.

b) if your goal is to further train the model, so it works on many tasks, RL helps with exploring new paths and training the model further. Constrained grammars help with inference, but the model doesn't "learn" anything. With RL you can also have many reward functions at the same time. Say one that rewards good syntax, one that rewards "closing" all the functions so tree-sitter doesn't complain, and one that rewards 0 errors from the compiler. The model gets to train on all 3 at the same time.

rationably•1h ago
Do you plan to release Sweep 3B/7B on HF?
_ache_•1h ago
Yeap, the two seems like game changer. For now, I'm using "Qwen2.5-Coder-7B". Sweep 1.5B is "just" 12 % point better than Qwen2.5-Coder, but Sweep 7B is 25% point better.
_ache_•1h ago
It's good. The blog post about it is very interesting. I hope, a plugin for neovim will be made soon.

https://blog.sweep.dev/posts/oss-next-edit

_boffin_•1h ago
Followed your work since the beginning and used it for inspiration for some cool demos on self-healing web scrapers. fascinating to see the transition from original concept to producing models. cool stuff.
whimsicalism•1h ago
Very interesting - and cool to read about the development process. I'd love to hear more about how genetic algorithm worked here.

I wonder whether we are perhaps the point of usefulness of 'next edit' code development in 2026 though.

dainiusse•1h ago
Any easy way to try on vscode?
esquire_900•43m ago
Surprising how badly Jetbrains implemented AI. Apparently to such an extent that even after multiple years of LLM's someone felt confident enough to build a company that can do better.

This looks really neat, interesting technical writeup as well!

h33t-l4x0r•22m ago
It sounds like you might be killing Zed's ability to monetize, am I misunderstanding that?
kleiba•20m ago
Very cool!

I understand that the 1.5B is small enough to run locally... but does it actually in the Sweep AI Jetbrains plugin? That is, if I install the plugin, will I download the model automatically and the plugin doesn't phone home?

moelf•2m ago
what do people use for Neovim to integrate these models for tab-completion level of stuff. (i.e. non agentic/vibe coding)

Show HN: Sweep, Open-weights 1.5B model for next-edit autocomplete

https://huggingface.co/sweepai/sweep-next-edit-1.5B
158•williamzeng0•7h ago•23 comments

Show HN: ChartGPU – WebGPU-powered charting library (1M points at 60fps)

https://github.com/ChartGPU/ChartGPU
565•huntergemmer•16h ago•162 comments

Show HN: High speed graphics rendering research with tinygrad/tinyJIT

https://github.com/quantbagel/gtinygrad
20•quantbagel•3h ago•6 comments

Show HN: Rails UI

https://railsui.com/
153•justalever•12h ago•83 comments

Show HN: RatatuiRuby wraps Rust Ratatui as a RubyGem – TUIs with the joy of Ruby

https://www.ratatui-ruby.dev/
111•Kerrick•4d ago•18 comments

Show HN: Differentiable Quantum Chemistry

https://github.com/lowdanie/hartree-fock-solver
24•lowdanie•4d ago•4 comments

Show HN: Open-source certificate from GitHub activity

https://certificate.brendonmatos.com
30•brendonmatos•3d ago•7 comments

Show HN: MCP server for searching and retrieving 200k icons

https://github.com/better-auth/better-icons
2•bekacru•1h ago•0 comments

Show HN: Dotenv Mask Editor: No more embarrassing screen leaks of your .env

https://marketplace.visualstudio.com/items?itemName=xinbenlv.dotenv-mask-editor
15•xinbenlv•7h ago•16 comments

Show HN: yolo-cage – AI coding agents that can't exfiltrate secrets

https://github.com/borenstein/yolo-cage
51•borenstein•16h ago•65 comments

Show HN: Retain – A unified knowledge base for all your AI coding conversations

https://github.com/BayramAnnakov/retain
32•Bayram•11h ago•12 comments

Show HN: Yashiki – A tiling window manager for macOS in Rust, inspired by River

https://github.com/typester/yashiki
16•typester•3d ago•10 comments

Show HN: ERCOT Texas power grid dashboard

https://txryan.com/ercot
2•nexuszero•2h ago•0 comments

Show HN: See the carbon impact of your cloud as you code

https://dashboard.infracost.io/
61•hkh•16h ago•22 comments

Show HN: Mastra 1.0, open-source JavaScript agent framework from the Gatsby devs

https://github.com/mastra-ai/mastra
208•calcsam•1d ago•69 comments

Show HN: Company hiring trends and insights from job postings

https://jobswithgpt.com/company-profiles/
43•sp1982•13h ago•6 comments

Show HN: Semantic search engine for Studio Ghibli movie

https://ghibli-search.anini.workers.dev/
24•aninibread•17h ago•7 comments

Show HN: Laptop Stickers – cheap individual short run stickers

https://laptopstickers.store/
6•decryption•3h ago•1 comments

Show HN: SpeechOS – Wispr Flow-inspired voice input for any web app

https://www.speechos.ai/
12•gangster_dave•15h ago•5 comments

Show HN: PicoFlow – a tiny DSL-style Python library for LLM agent workflows

5•shijizhi_1919•16h ago•0 comments

Show HN: Hyve – Parallel isolated workspaces for coding agents, multi-repo dev

9•eladkishon•23h ago•1 comments

Show HN: Agent Skills Leaderboard

https://skills.sh
130•andrewqu•1d ago•41 comments

Show HN: I built a chess explorer that explains strategy instead of just stats

https://www.atlaschess.me/
11•Ahmad_shuja•15h ago•5 comments

Show HN: Automatically build sales playbook. For founders doing sales

7•Mrakermo•11h ago•1 comments

Show HN: TopicRadar – Track trending topics across HN, GitHub, ArXiv, and more

https://apify.com/mick-johnson/topic-radar
35•MickolasJae•1d ago•9 comments

Show HN Guidelines

https://news.ycombinator.com/yli.html
2•cjbarber•8h ago•0 comments

Show HN: Ocrbase – pdf → .md/.json document OCR and structured extraction API

https://github.com/majcheradam/ocrbase
96•adammajcher•1d ago•34 comments

Show HN: I'm eating at all the phở restaurants in Portland, at least twice

https://pho.curtisbarnard.com/
3•oregoncurtis•9h ago•4 comments

Show HN: I built an AI coach for introverted leaders

https://www.leadquiet.com/landing
2•chux52•10h ago•0 comments

Show HN: Sornic – Turn any article into a podcast in 10 seconds

https://sornic.com
2•digi_wares•10h ago•1 comments