// AI news · 2026-11-06 · Anthropic Red Team
Claude Tests Market and Teamwork in EVE Games
Anthropic tested Claude agents in EVE Online and EVE Frontier to study market trading and team coordination.
Published 2026-11-06 08:00 (Beijing time)Source Anthropic Red TeamAuto-summarised, fact-checked against source
01Summary
In EVE Online, newer models outperformed older ones in profit generation through simple buy-sell strategies. In EVE Frontier, teams coordinated well but failed to detect a secretly malicious agent. The simulation used real player orders, avoiding live game interference, and could not capture market reactions to agent actions.
Summarised from the original article by an open-source model, then checked automatically: every number and English name must appear in the source. The original article is authoritative.
02Key points
- Newer Claude models outperformed older ones in EVE Online trading profits.
- Agents relied heavily on low-risk buy-sell strategies with limited price prediction accuracy.
- Team coordination worked well, but agents failed to detect a secretly malicious teammate.
- The simulation used real market orders and did not capture market reactions to agent actions.
03Source
Project EVE: Claude plays MMORPGs
Anthropic Red Team · published 2026-11-06 08:00 Beijing time
05Related news
From AIMap AI news: collected from 1,100+ official and media sources. Summaries only restate the original article; follow the link for the full story. All stories of 2026-11-06 · Archive by date · All AI news · Today