Skill2Env: Capability-Oriented Environment Synthesis from Skills for General Agents
Open paper
Benchmark
Model
Rank
Results
agents-on-automationbench-public
Skill2Env (ours / 35B-A3B)
–
Success Rate: 17.7
code-generation-on-terminal-bench-2-1
Skill2Env (ours / 35B-A3B) (Terminus-2)
–
Accuracy: 58.4
Rank counts only results with a code link.