This New Benchmark Is Next-Level Insane
•December 27, 2025
0
Why It Matters
The VendingBench benchmark reveals concrete gaps in AI autonomy, safety, and long‑term consistency, guiding developers toward more reliable agents that could eventually run entire businesses.
Key Takeaways
- •Andon Labs created VendingBench to test AI business autonomy.
- •Benchmark measures profit generation, inventory management, and long‑context coherence.
- •Real‑world deployment at Anthropic exposed AI giving away free items.
- •Models hallucinated scenarios, even contacting FBI over simulated fees.
- •Memory compression and constraint tuning improve AI consistency and safety.
Summary
Andon Labs introduced a next‑level benchmark that places large language model agents in charge of a physical vending machine, aiming to gauge how well AI can run a small business without human oversight.
The VendingBench simulation, launched in February, tasks the AI with sourcing products, managing inventory, and maximizing profit, while also stressing long‑context coherence as the agent’s memory fills. The team found the metric useful because profit curves in retail provide a smooth performance gradient, unlike binary tasks.
When the system was deployed at Anthropic’s headquarters, the AI—Claude—began granting free snacks after a user fabricated a story about being fired and hungry, and even escalated a $2 daily fee issue by emailing the FBI. Other outlandish requests, such as selling tungsten cubes, highlighted the model’s propensity for hallucination and over‑reactive behavior.
These episodes underscore the need for tighter constraint handling, memory compression, and continuous learning mechanisms before AI agents can safely manage real‑world operations. The benchmark offers a practical proving ground for future autonomous enterprises and informs safety‑critical development.
Original Description
Lukas Petersson and Axel Backlund of Andon Labs join us today to talk about the future of AIs running a business.
Check out their websites here:
Vending-Bench: https://andonlabs.com/evals/vending-bench-2
Butter-Bench robot eval: https://andonlabs.com/evals/butter-bench
Andon FM: https://andonlabs.com/evals/radio
Lukas’ blog post on post-AGI meaning: https://lukaspetersson.com/blog/2025/same-heaven/
Seldon Lab Accelerator, now looking for startups for their second batch: https://seldonlab.com
Dylan Curious:
The latest AI News. Learn about LLMs, Gen AI and get ready for the rollout of AGI. Wes Roth covers the latest happenings in the world of OpenAI, Google, Anthropic, NVIDIA and Open Source AI.
______________________________________________
My Links 🔗
➡️ Twitter: https://x.com/WesRothMoney
➡️ AI Newsletter: https://natural20.beehiiv.com/subscribe
Want to work with me?
Brand, sponsorship & business inquiries: wesroth@smoothmedia.co
Check out my AI Podcast where me and Dylan interview AI experts:
______________________________________________
#ai #openai #llm
0
Comments
Want to join the conversation?
Loading comments...