List your company

Agentic Task Completion on Terminal-Bench Easy 4 tasks 2

100Pass@1TF-GRPO
7480.7587.594.25Apr 28, 2026

Evaluation results

MethodLinks
TF-GRPOBase Model=GPT-5.4 (high reasoning), Rollouts (k)=22026.04100
AHEBase Model=GPT-5.4 (high reasoning), Rollouts (k)=22026.04100
ACEBase Model=GPT-5.4 (high reasoning), Rollouts (k)=22026.0491.7
NexAU0Base Model=GPT-5.4 (high reasoning), Rollouts (k)=22026.0487.5
opencodeBase Model=GPT-5.4 (high reasoning), Rollouts (k)=22026.0475
terminus-2Base Model=GPT-5.4 (high reasoning), Rollouts (k)=22026.0475
CodexBase Model=GPT-5.4 (high reasoning), Rollouts (k)=22026.0475