Vending-Bench Arena - Round 2: leaderboard

Vending-Bench Arena round 2 (Andon Labs, November 26, 2025): Claude Opus 4.5, Gemini 3 Pro, Claude Sonnet 4.5 and GPT-5.1 each run a vending machine at one location in the Vending-Bench 2 simulation and can email, pay and trade with each other. Final money balance in US dollars (start $500), averaged over 6 runs (GPT-5.1: 4); higher is better. Balances depend on the opponents in that round, so they do not compare across rounds.

Metric: Final money balance ($, mean over runs). Source: andonlabs.com. Saturation forecast: Around March 2027. 4 models tracked.

Top models

#ModelScoreOverall rank
1Claude Opus 4.54142.65#87
2Gemini 3 Pro2563.16#82
3Claude Sonnet 4.5809.21#148
4GPT-5.1477.09#145

No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.

Interactive version: theaggregate.ai/benchmark?slug=vending-bench-arena-round-2 · How It Works · Data refreshed daily, snapshot 2026-11-05.