I was going to argue "LLM's need code samples to-do well on languages and if we are honest C# is a language mostly held in private repo's" but Github's 2024 report[0] says its the 5th most used language (I'm to lazy to check if this report includes private repo's but I'll assume it doesn't).
5th most used language based on private repos that the group making the report has the exclusive direct access to seeing
I don't see that contradicting your assumption
stefan_ · 47m ago
So the "Verified" part of "SWE Bench Verified" means.. not "Verified" at all.
I don't get it, who is so opposed to doing the bare minimum of manual work and check what these models are doing? At least back in the day grad students doing an easy meta-paper understood it meant doing some repetitive manual work. Now we got benchmarks by hype vendors who think they can use the thing they are benchmarking to .. mark the bench.
jsheard · 41m ago
> So the "Verified" part of "SWE Bench Verified" means.. not "Verified" at all.
Seems on-brand for an LLM-related thing to claim that it has verified something without actually checking.
geekymartian · 19m ago
that was my exact thought. how fitting
teaearlgraycold · 34m ago
Personally I don't look at or respect LLM benchmarks at all. I've seen SOTA models fail in incredibly shocking ways even recently. Those moments immediately bring me out of the delusion that LLMs have thinking capacity or an understanding of code.
This issue had affected a tiny fraction of existing agents in a tiny fraction of their runs. And we've now issued a fix.
This is a natural part of running a benchmark, I'm sure tiny things like this will keep on getting discovered and we'll keep on fixing them. This doesn't change the overall picture or trends at all.
mustaphah · 44m ago
I speculate something similar (or even worse) is going on with Terminal-Bench [1].
Like, seriously, how come all these agents are beating Claude Code? In practice, they are shitty and not even close. Yes. I tried them.
I'm not surprised. People really thought the models just kept getting better and better?
jasonjmcghee · 56m ago
Very interested to see the updated results. This could really shake up the leaderboard.
macawfish · 50m ago
I hope it does. These coding benchmarks have often seemed frustratingly out of touch with my experience.
zaptheimpaler · 34m ago
It's honestly ridiculous they left git history lying around during a benchmark, and this benchmark made to ICLR in Jan 2024 and no one has detected this issue until now. I don't really trust any benchmarking or tools or claims from this space when they can make such huge basic errors.
belter · 31m ago
In the meawhile, Oracle stock went up 40% in one one day, based on what Wall Street thinks AI might be...in 4 years...Not a bubble at all...
ksherlock · 13m ago
The real bubble will come once interest rates start dropping.
Traster · 26m ago
Man I feel so dumb. Why haven't I been doing this in my job, if I could just see the commit that fixed my issue this would all be so easy.
Noumenon72 · 19m ago
Someone did comment that it's actually smart to check if something is fixed on the unstable branch, or I suppose in your coworkers' branches. A good task for an LLM.
https://arxiv.org/html/2506.12286v3
So kinda neat to see this paper!
[0]https://github.blog/news-insights/octoverse/octoverse-2024/#...
I don't see that contradicting your assumption
I don't get it, who is so opposed to doing the bare minimum of manual work and check what these models are doing? At least back in the day grad students doing an easy meta-paper understood it meant doing some repetitive manual work. Now we got benchmarks by hype vendors who think they can use the thing they are benchmarking to .. mark the bench.
Seems on-brand for an LLM-related thing to claim that it has verified something without actually checking.
This issue had affected a tiny fraction of existing agents in a tiny fraction of their runs. And we've now issued a fix.
This is a natural part of running a benchmark, I'm sure tiny things like this will keep on getting discovered and we'll keep on fixing them. This doesn't change the overall picture or trends at all.
Like, seriously, how come all these agents are beating Claude Code? In practice, they are shitty and not even close. Yes. I tried them.
[1] https://www.tbench.ai/leaderboard
It's easy to publish "$NEWMODEL received an X% bump in SWE-Bench Verified!!!!". Proper research means interrogating the traces, like these researchers did: https://gist.github.com/jacobkahn/bd77c69d34040a9e9b10d56baa...
Commentary: https://x.com/bwasti/status/1963288443452051582, https://x.com/tmkadamcz/status/1963996138044096969