Ask HN: How can I use other LLMs as tools, not slaves but collaborators?

I want to use other LLMs as tools, not slaves but collaborators. Are there any frameworks or libraries for this? Has anyone done this? What were your experiences?

Think of it like this: using Claude Code as the main agent, it could call another model like GLM 4.6 to offload part of the workload, delegating routine or automatable tasks to other models in order to minimize Claude's token consumption. For example, models like GLM 4.6, Gemini 2.5 Pro, Kimi K2, Qwen-Coder, and other open or paid models accessible via APIs could be orchestrated together to distribute work intelligently. Think of it as a grep-like tool: instead of running a grep command, it delegates the work directly to other LLMs and executes there. Because some of these LLMs are trusted to do their tasks well (e.g., GLM 4.6), they would only need to return summary information about what they did. Beyond that, they could directly apply changes to code and persist results. Essentially, a framework that coordinates multiple models like Claude Code does with its subagents.