gpt-6IMAGE
GPT-6 Astra 在 6 种 Agent 框架上的 29 项任务实测
Composio 把 Astra 放进 Codex、Claude Code 等 6 个 agent 框架跑 29 个难题，成功率相近，但失败时 token 消耗高出 3–5 倍。
Composio
@composio2026.09.17收录原推发布于 2026.09.17
Composio @composio原推发布于 2026.09.17
We ran GPT-6 Astra across 6 agent harnesses (Codex, Claude Code, OpenCode, Hermes Agent, Pi Agent, Command Code) on 29 challenging agentic tasks. Most harnesses succeeded at similar rates. But when they failed, they used 3–5x as many tokens, depending on the harness. 🧵🧵🧵
x.com/composio/status/2100308380980068538采集于 2026.09.17 10:17