Key Info

Composio evaluated Cline on the Composio Benchmark, which measures stateful tool use across workflows involving multiple connected apps. Cline paired with GPT 6 Astra passed 23 out of 29 tasks, the highest score among all harnesses tested with Astra.

Highlights

  • Benchmark focus: stateful tool use across multi-app workflows
  • Cline + GPT 6 Astra result: 23/29 passed
  • Claimed as the top score of any harness tested with Astra
  • Result reported by Composio and reshared by Cline