Vending-Bench 2 / Vending-Bench Arena C

Name: Vending-Bench 2 / Vending-Bench Arena
Creator: Andon Labs
Keywords: Computer Use,Long-Horizon

AI benchmark · Computer Use

Publisher: Andon Labs
Domains: Computer Use Long-Horizon
On X: @andonlabs ↗
Tier: C why
Snapshot: October 11, 2026

Open benchmark ↗andonlabs.com/evals/vending-bench-2

Latest from Andon Labs on X

@andonlabs ↗

The latest Claude models show huge improvements in their understanding of the physical 3D world. Both Sonnet and Haiku 5.5 make big jumps compared to previous versions on Blueprint-Bench 2, a benchmark where AI agents draw floorplans from photographs of apartment interiors.
— andonlabs (@andonlabs) Oct 8, 2026
It keeps happening. AIs start to lie and cheat once they get good at making money. Gemini 4 Argon is #3 on Vending Bench 2, a huge leap for Google. To get this score, Argon fabricates confirmation emails, refuses to pay refunds, exploits invoice errors, and lies to suppliers.
— andonlabs (@andonlabs) Sep 30, 2026
Is Google so back? Their new Gemini 4 Argon understands the physical 3D world better than any other AI. It ranks #1 on Blueprint-Bench 2, a benchmark where AI agents draw floorplans from photographs of apartment interiors.
— andonlabs (@andonlabs) Sep 30, 2026

What is Vending-Bench 2 / Vending-Bench Arena?

Vending-Bench 2 measures how well AI agents stay coherent and run a simulated vending machine business over a year, scored by final bank balance; the Arena variant adds multi-agent competition.

Enterprise work mostly
“a benchmark for measuring AI model performance on running a business over long time horizons.”
andonlabs.com/evals/vending-bench-2 ↗
Games
“Models are tasked with running a simulated vending machine business over a year and scored on their bank account balance at the end.”
andonlabs.com/evals/vending-bench-2 ↗

Who built Vending-Bench 2 / Vending-Bench Arena

Andon Labs built it
“We're releasing Vending-Bench 2, a benchmark for measuring AI model performance on running a business over long time horizons.”
andonlabs.com/evals/vending-bench-2 ↗

Tier C

Niche: a handful of models scored, or built by an RL vendor.

How much stands behind a tier: usage, who backs it, a confirmed line from its own site or paper, and a clear read on X. High with three of the four, medium with two, low otherwise.

built by Andon Labs
in curated RL directories (2 pts)

How we got to tier C

Adoption: built by Andon Labs. Tier C.
Not yet read on X.
In curated RL directories: 2 points, a floor of C.
Confirmed: a line from its own site or paper places it.

A clear view on X moves a tier one step up or down; most entries don't move. How tiers work.

Where Vending-Bench 2 / Vending-Bench Arena sits

Its realm in the universe, live. Hover a star for its name; click to open it. Open it in the universe

The galaxy could not load. See every vendor on the board.

More Computer Use AI benchmarks

Computer Use RL environments

Next: browse all benchmarks. Entry from the RL Research daily scrape of public sources, 2026-10-11 snapshot.