Randomly switching between LMs at every step boosts SWE-bench score
Hello from the SWE-bench/SWE-agent team at Princeton/Stanford. When we created the SWE-bench benchmark in 2023 from hundreds of real-life GitHub issues/pull requests, the highest score was just a couple of percent. The tasks were so challenging for LMs, that most people didn't even want to work on them. Half a year later, SWE-agent showed that the early 2024 LMs were actually good enough to resolve up to 20% of the GitHub issues in the benchmark. This kicked off a whole wave of coding agents. Back then, developing agents was all about working around tons of silly behavior from…
Randomly switching between LMs at every step boosts SWE-bench score launched on August 20, 2025 with 5 votes and more than 75% of that year's launches. The site was not reachable when checked in September 2026.
What it does
In the maker’s words, at launch
What if your agent uses a different LM at every turn? We let mini-SWE-agent randomly switch between GPT-5 and Sonnet 4 and it scored higher on SWE-bench than with either model separately. GPT-5 by itself gets 65.0%, Sonnet 4 64.8%, but randomly switching at every step gets us 67.2% This result came pretty surprising…
Does a similar job
all alternatives →
New Benchmark from SWE-bench team is 0% solvedMay 2026 · programbench.com · ▲24ProgramBench evaluates whether language models can rebuild programs from scratch.
BulletAug 2026 · codewithbullet.com · ▲240Hi HN, excited to be sharing this with you guys today. TL;DR: Bullet is a fast coding agent. At our company, we were burning hours waiting on agent runs. The models weren’t the problem. The agent loops around them were slow (unnecessary planning, oversized context, repeated searches, and serial tool calls). So we built Bullet as a passion project, the goal being to help you iterate faster, without giving up quality: - It automatically selects the model and reasoning level for each prompt. - It runs independent searches, reads, and commands in parallel. - It uses targeted code search instead…
OSS Agent I built topped the TerminalBench on Gemini-3-flash-previewApr 2026 · github.com · ▲393Scored 65.2% vs google's official 47.8%, and the existing top closed source model Junie CLI's 64.3%. Since there are a lot of reports of deliberate cheating on TerminalBench 2.0 lately (https://debugml.github.io/cheating-agents/), I would like to also clarify a few things 1. Absolutely no {agents/skills}.md files were inserted at any point. No cheating mechanisms whatsoever 2. The cli agent was run in leaderboard compliant way (no modification of resources or timeouts) 3. The full terminal bench run was done using the fully open source version of the agent, no…
Coarena by CoastyAug 2026 · coarena.ai · ▲111We built a computer use agent once and got it benchmarked at the 2nd position in the biggest benchmark for this field but in the process, we realized one thing, how outdated and inefficient these benchmarks are and how they don't truly capture how these tasks went. So we thought a community led benchmarking arena would be the way to go about. https://coarena.ai/. You just have to put in a task and two side by side windows with hidden models will show up and do the same task and at the end you vote which one did better(or none). We have a leaderboard where we rank all the…
Muscle-Mem, a behavior cache for AI agents2025 · github.com · ▲226Hi HN! Erik here from Pig.dev, and today I'd like to share a new project we've just open sourced: Muscle Mem is an SDK that records your agent's tool-calling patterns as it solves tasks, and will deterministically replay those learned trajectories whenever the task is encountered again, falling back to agent mode if edge cases are detected. Like a JIT compiler, for behaviors. At Pig, we built computer-use agents for automating legacy Windows applications (healthcare, lending, manufacturing, etc). A recurring theme we ran into was that businesses already had RPA (pure-software scripts), and…
A registry of agent benchmarks (including many OSS agent trajectories)2024 · explorer.invariantlabs.ai · ▲6If you're interested in exploring what LLM-based agent systems these days actually do to solve certain benchmarks such as SWEBench or WebArena, we created a small leaderboard with our team, that allows to view a lot of public and OSS agent results including all the runtime traces (the step-by-step reasoning behind the scenes). Looking at traces is actually quite interesting, as they reveal a lot about the inner working and shortcomings of current agent system, e.g. see https://explorer.invariantlabs.ai/u/invariant/webarena--SteP... for an example trace.
More ai this month
the category →
Paintings by AI models: programs run through a simulation of oil paint.
AI · 21d ago · stillwet.art
PixVerse R2▲381A real-time world model you can explore and change
AI · 30d ago · world.pixverse.video
Source: Show HN post ↗ · launch text quoted in part
Launched alongside, August 2025
the whole month →
I built the world's most impractical 1000-pixel display and anyone in the world can draw on it. It draws a single pixel at a time and takes 30-60 minutes to complete a single image. Anyone can participate in the project by voting for the next image to be drawn, and submitting images. https://kilopx.com/
Work · 2025 · benholmen.com
Kitten TTS is an open-source series of tiny and expressive text-to-speech models for on-device applications. We are excited to launch a preview of our smallest model, which is less than 25 MB. This model has 15M parameters. This release supports English text-to-speech applications in eight voices: four male and four female. The model is quantized to int8 + fp16, and it uses onnx for runtime. The model is designed to run literally anywhere eg. raspberry pi, low-end smartphones, wearables, browsers etc. No GPU required! We're releasing this to give early users a sense of the latency and voices…
AI · 2025 · github.com
I was wondering how I can arrange objects along a spherical helix path, and read some articles on it. I ended up learning about parametric equations again, and make this visualization to document what I learned: https://visualrambling.space/moving-objects-in-3d/ feel free to visit and let me know what you think!
Life & fun · 2025 · visualrambling.space
For HTML Day 2025 [1], I made a web service that displays the current sky at your approximate location as a CSS gradient. Colours are simulated on-demand using atmospheric absorption and scattering coefficients. Updates every minute, without the use of client-side JavaScript. Source code and additional information is available on GitHub: https://github.com/dnlzro/horizon [1] https://html.energy/html-day/2025/index.html
Dev tools · 2025 · sky.dlazaro.ca