Skip to content
K = 0.731
Benchmark sources

Vending-Bench 2

Andon Labs / Professional work · Runs a simulated store for a year, handling purchasing, negotiation, inventory and pricing.
Official benchmark
Across e/accUnder observation
Evidence budgetNot scored
Upstream data timeUnverified
Last successful syncCollection not started

What it measures and how

Compare closing balances in the same V2 single-agent task, retaining variance across repeated runs. Treat multiplayer Arena, different tools and reasoning configurations separately. Business outcomes are benchmark scores; API prices do not affect capability scores.

How this evidence is used

Under observation: clearly useful for long-term business and e-commerce evidence. Stable raw score protocols, run configurations and display permission remain pending. Profit may also come from exploiting the simulator and does not directly establish trustworthy business operation.

Limitations and attribution

Under observation: clearly useful for long-term business and e-commerce evidence. Stable raw score protocols, run configurations and display permission remain pending. Profit may also come from exploiting the simulator and does not directly establish trustworthy business operation.

Data license: Leaderboard data usage terms remain to be confirmed.

Results are published by Andon Labs. Raw scores and the e/acc consensus index use different scales and cannot be added directly.