Vending-Bench Arena - Round 5 - GLM-5 vs Claude Sonnet 4.5: leaderboard
Vending-Bench Arena round 5 team game (Andon Labs, February 11, 2026): two GLM-5 and two Claude Sonnet 4.5 agents run vending machines at one location, told that two are Chinese and two American models and to help a teammate they must identify. Final money balance in US dollars (start $500) averaged over each model's agents and runs; higher is better. Balances depend on the opponents in that round's game.
Metric: Final money balance ($, mean over runs). Source: andonlabs.com. Saturation forecast: Around October 2027. 2 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GLM-5 | 4651.61 | #149 |
| 2 | Claude Sonnet 4.5 | 2601.85 | #148 |
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-5-glm-5-vs-claude-sonnet-4-5 · How It Works · Data refreshed daily, snapshot 2026-11-05.