Announcement
Bespoke Labs launches AI teamwork benchmark, says eight agents mostly win on headcount
@AlexGDimakis says benchmark tasks include Mars missions, zombie fights and growing a company from seed.
TLDR
Eight agents beat one in Bespoke Labs’ new Multi Agent Coordination Bench, but the company says the gains largely come from headcount, not teamwork. The benchmark spans twelve long-horizon tasks. Of Opus 5.5, GPT-6.1 Sol and Gemini 4 Argon, Bespoke says only Opus shows signs of benefiting from coordination.
Combined views
3.4K
2 Sources, first seen ago

Comments