Benchmark
Vending Bench
Reported in simulated dollars, not a percentage, and run over long horizons where a single early failure dominates the total.
- Released
- 2025-02-18
- Categories
- agent
- Openness
- unknown
- Source
- model_reports
- Reported scores
- 2
Reported scores
model_reports
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| GLM-5 | Z.ai | 4432.12 | 2026-02-11 | evidence |
| GLM-5.1 | Z.ai | 5634.41 | 2026-04-03 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.