Search papers, labs, and topics across Lattice.
1
0
2
5
Despite the promise of advanced AI agents, even the best performers struggle, with Codex GPT-5.5 only achieving a 42% success rate in complex healthcare tasks.