We had a great time presenting at #PyHEP2022 where we talked about our goals, methodology, curriculum as well as how to become a part of our efforts! We invite you to join us in building a better future for HEP! Find out more about the talk here: indico.cern.ch/event/1150631/…

It's been just 10 days since we launched SWE-agent but we already have 1.5k people in our Discord and lots of contributors on GitHub. We've been making the agent easier to use and there are lots more exciting updates coming soon, including a web UI! Join us :)

You can now apply SWE-agent to any local repository and use any text file as the input issue instead of having to use GitHub repos/issues. Lots of people were asking for this! More information in the latest release notes: github.com/princeton-nlp/…

Run SWE-agent to solve GitHub issues completely in-browser, no installation required. Try it now: codespaces.new/princeton-nlp/…

The SWE-agent preprint has finally landed! Check it out at swe-agent.com/paper.pdf

We’re excited to launch the SWE-agent web UI! Specify a bug, press start and watch SWE-agent do the magic ✨ github.com/princeton-nlp/…

📊Want to add your AI Coder / Language Model to the SWE-bench leaderboard (swebench.com)? SWE-bench official submission process is now available! github.com/swe-bench/expe… Simply add your model's predictions, execution logs, and agent agent trajectories to the repo.

SWE-agent release v0.5.0 brings extensive documentation/tutorials and speedup by caching environment setups to skip installation steps. We also reformatted our codebase; see our docs for rebasing your fork. princeton-nlp.github.io/SWE-agent/inst…

SWE-agent just got 2x faster! We also expanded our tutorial with more details on speedup and environment setup! Full changelog: princeton-nlp.github.io/SWE-agent/inst…

The GenAI Collective had the privilege of hosting the esteemed Princeton researchers behind SWE-bench and SWE-agent at our first ever NYC research meetup! Huge shoutout to Ofir Press John Yang carlos and Kilian Lieret for informative talks and hanging with our community The

If you want to start working with and extending SWE-agent, Kilian Lieret just wrote this detailed overview of the architecture of SWE-agent: princeton-nlp.github.io/SWE-agent/back…

If you want to use SWE-agent for LeetCode-style problems, we just made that a bit easier: princeton-nlp.github.io/SWE-agent/usag…

We're super excited to see how people use SWE-agent to solve super tough programming challenges So we built a SWE-agent starter kit for the NeurIPS Meta Hacker Cup AI track :) Good luck!

We are live with the NeurIPS AI Hacker Cup Challenge, featuring insightful lectures on SWEAgent with Kilian Lieret. Stream now: youtube.com/live/d9gcXpiiD…

Cool new benchmark from Ben Bogin et al that checks whether agents can run experiments in repos from research papers. They adapt SWE-agent for this task and it gets impressive results! Super happy to see our system being used for new & creative tasks.

We're launching SWE-bench Multimodal to eval agents' ability to solve visual GitHub issues. - 617 *brand new* tasks from 17 JavaScript repos - Each task has an image! Existing agents struggle here! We present SWE-agent Multimodal to remedy some issues Led w/ carlos 🧵

Can agents solve front-end engineering issues with images? *SWE-bench multimodal* turns more than 600 real JS GitHub into a benchmark. None of the other agents we tested generalized well to this challenge, so we added multimodal capabilities to SWE-agent!

Running agents with execution feedback is hard! Next week, we will release 🦖SWE-ReX, an agent-agnostic middleware that handles sandboxing and managing shell sessions for you - locally or in the cloud! We’ll launch with SWE-agent 1.0.0, a significant refactor on top of SWE-ReX.