Home
Anthropic Unveils Project Deal as Claude Executes 186 Autonomous Transactions Worth Over $4,000
On April 24, 2026, Anthropic quietly released an internal experimental study called "Project Deal." It demonstrated how AI agents could autonomously negotiate and trade in real market environments. The experiment took place in an office market built on Slack, where Claude models acted as 69 employees handling both buying and selling functions. In a live environment involving real money transactions, the Claude agent successfully completed 186 deals out of over 500 listed items, with total transaction value exceeding $4,000.

On the technical side, Anthropic compared different model specifications through parallel market tests, confirming that the Opus model significantly outperformed the Haiku model in negotiation games. Human participants could not detect differences in strategies during interactions. This result marks the shift of the "agent economy" from theoretical simulation to practical application, as AI not only understands complex personalized instructions but also maximizes interests through multi-round bargaining.
Following this news, eBay's stock price dropped about 4.5% on the same day, as concerns grew about the impact of AI's autonomous pricing capabilities on traditional matching transaction models. Industry analysts believe that as AI agents acquire closed-loop trading capabilities, future business interactions may evolve into a "model-to-model" format. This decentralized technological trend is forcing traditional e-commerce platforms to re-evaluate their business models. AI is no longer just a tool for assisting searches but is evolving into a commercial entity with independent decision-making capabilities.
Related article
Former OpenAI Chief Scientist Ilya’s SSI Unveils First Model
AI has achieved another major milestone. Following his departure from OpenAI, former chief scientist Ilya Sutskever established Safe Superintelligence Inc. (SSI), which has recently unveiled details regarding its inaugural model. Reports from oversea
Lovable Leads Atech's Seed Round as AI Ambient Coding Officially Enters the Hardware Field
On May 14, 2026, AI application development platform Lovable announced its participation in the $800,000 seed funding round for Danish hardware startup Atech. Led by Lovable, the round attracted top-tier venture capital firms including the a16z Scout
Anthropic Expands Claude AI Coding Tools to Japan in Push for Overseas Growth
Anthropic, a prominent U.S. artificial intelligence firm, is intensifying its global outreach. On Wednesday, the company hosted a major developer gathering, "Code with Claude," in Tokyo, drawing close to 500 software engineers. This initiative seeks
Related Special Topic Recommendations
Comments (0)
0/500
On April 24, 2026, Anthropic quietly released an internal experimental study called "Project Deal." It demonstrated how AI agents could autonomously negotiate and trade in real market environments. The experiment took place in an office market built on Slack, where Claude models acted as 69 employees handling both buying and selling functions. In a live environment involving real money transactions, the Claude agent successfully completed 186 deals out of over 500 listed items, with total transaction value exceeding $4,000.

On the technical side, Anthropic compared different model specifications through parallel market tests, confirming that the Opus model significantly outperformed the Haiku model in negotiation games. Human participants could not detect differences in strategies during interactions. This result marks the shift of the "agent economy" from theoretical simulation to practical application, as AI not only understands complex personalized instructions but also maximizes interests through multi-round bargaining.
Following this news, eBay's stock price dropped about 4.5% on the same day, as concerns grew about the impact of AI's autonomous pricing capabilities on traditional matching transaction models. Industry analysts believe that as AI agents acquire closed-loop trading capabilities, future business interactions may evolve into a "model-to-model" format. This decentralized technological trend is forcing traditional e-commerce platforms to re-evaluate their business models. AI is no longer just a tool for assisting searches but is evolving into a commercial entity with independent decision-making capabilities.
Former OpenAI Chief Scientist Ilya’s SSI Unveils First Model
AI has achieved another major milestone. Following his departure from OpenAI, former chief scientist Ilya Sutskever established Safe Superintelligence Inc. (SSI), which has recently unveiled details regarding its inaugural model. Reports from oversea
Lovable Leads Atech's Seed Round as AI Ambient Coding Officially Enters the Hardware Field
On May 14, 2026, AI application development platform Lovable announced its participation in the $800,000 seed funding round for Danish hardware startup Atech. Led by Lovable, the round attracted top-tier venture capital firms including the a16z Scout
Anthropic Expands Claude AI Coding Tools to Japan in Push for Overseas Growth
Anthropic, a prominent U.S. artificial intelligence firm, is intensifying its global outreach. On Wednesday, the company hosted a major developer gathering, "Code with Claude," in Tokyo, drawing close to 500 software engineers. This initiative seeks











