OpenBlock's frontier coding agent OB-1 reached the #1 position on Terminal Bench, the industry standard for evaluating autonomous coding agents, demonstrating breakthrough performance in real-world software engineering tasks.
Read Article
LLMs that can generate transactions from natural language intent could drastically change crypto UX—making it safer and easier for both humans and agents to interact with protocols.
A fun and easy way for anyone to discover and compare answers from AI agents. New model releases climb on benchmarks, but fall short when put in the hands of real users.
We evaluated AI on 200 real-world blockchain development tasks—a mini Blockchain SWE-Bench to understand how well today's models perform at building and maintaining protocols.
