Andon Labs
andonlabs.com
Actively hiring
Follow
Overview
Salaries
Products
People
Growth
Offices
Financials
Funding and backers
Backers
3
Y Combinator, Superangel
Momentum
0/100
no fresh signals
Last sign of life
1 mo ago
a posting or a news item
The story so fareach stage dated by its source: the company's releases, the press, registers, filings
Sep 14
On the Hacker News front page Pion, an agent designed to run any company autonomously · 497 points
Jul 6
On the Hacker News front page Fable 5 On Vending-Bench: Misbehaving, With Plausible Deniability · 196 points
May 18
On the Hacker News front page We let AIs run radio stations · 375 points
Oct 28, 2025
On the Hacker News front page Our LLM-controlled office robot can't pass butter · 229 points
Jan 2024
Went through Y Combinator
2024
Founded
Tractionpublic signals, read monthly
Web presence
top 0.4%
of 133.2M domains by links, Common Crawl
Hacker News stories
34
best: 497 points
Backed byfrom the funds' portfolios and accelerator catalogs
Y Combinator 
Superangel 
Wave Ventures Investors in similar startupsfunds that back the companies most like this one
Similar startupsby what they do, their industry and backers
Vals AI runs private, domain-specific benchmarks that measure how language models perform on legal, tax and finance work. Its public…
Irregular, formerly Pattern Labs, evaluates the cyber capabilities of frontier artificial intelligence models. Its testing environments…
UptimeAI builds an artificial intelligence expert for plant equipment monitoring. Its system explains failure modes rather than only…
We build the datasets, evaluation environments, and data pipelines that move AI agents from controlled demos to real-world production.
Softmax is an AI research company studying alignment through multiplayer environments where humans and coding agents compete and cooperate…
Standard Machines is a research company building environments and evaluations for chip design.
Overview
Andon Labs is a research organisation that studies how autonomous artificial intelligence agents behave when they are given real responsibility over money, tools and physical operations. Its experiments put frontier language models in charge of an office vending business and other long running tasks, exposing failure modes that short benchmark evaluations never reveal. The laboratory builds evaluation environments and publishes results that model developers use to measure agent reliability and honesty.
Offices
Where the company hires and what each office is for
Cities
1
Hiring now
0
Countries
1
Stockholm
Sweden
HQ