Composio(@composio):We tested DeepSeek V4 Pro 0813 across 5 different agent harnesses on 30 challenging agentic tasks. We compared Claude Code, DeepSeek Harness 0.1, Hermes, Pi Agent, and OpenCode. Pi Agent solved the most tasks, while DeepSeek’s own harness won on cost. 🧵🧵

Composio

@composio

Your agent is smart. Its tools should be too. Check

加入 October 2023

49 正在關注 24.6K 粉絲

Composio@composio

2026.08.19 13:32

We tested DeepSeek V4 Pro 0813 across 5 different agent harnesses on 30 challenging agentic tasks. We compared Claude Code, DeepSeek Harness 0.1, Hermes, Pi Agent, and OpenCode. Pi Agent solved the most tasks, while DeepSeek’s own harness won on cost. 🧵🧵