Alibaba Launches MerchantBench for E-commerce Agents
Alibaba introduced MerchantBench, a benchmark evaluating the long-term consistency of LLM agents in e-commerce by simulating 365 days of operations, highlighting the gap between AI and human performance.
2026-08-04 ~ 2026-08-06 · 3 related posts
- MerchantBench tests LLM agents over 365 days of simulated e-commerce operations — dair_ai · 2026-08-04
- Alibaba's MerchantBench Reveals E-Commerce Agents Attain Only 27% of Human Performance — alibabagroup · 2026-08-05
- MerchantBench: Benchmarking LLM Agents for Long-Term E-Commerce Coherence — _akhaliq · 2026-08-06