GLM-5.2 Caveman Lite

GLM-5.2 Caveman Lite is presented by GreenPT as one of several lightweight variants in the GLM-5.2 lineup, positioned alongside caveman-ultra and other suffixes within the broader glm family cataloged on Models.dev. GreenPT's listing describes it as a text-in, text-out model that supports reasoning, tool calling, structured output, and temperature control, making it well-suited for agentic workflows, multi-step problem solving, and pipelines that require reliable function calling. Its open-weight status, as flagged in the GreenPT catalog, also means developers can self-host the model for offline or on-premises deployments where data residency matters, a practical advantage for teams integrating it into private stacks.

For long-context applications, the model is configured with a 1,000,000-token context window and a 131,072-token maximum output, which supports tasks such as extended document analysis, codebase reasoning, and retrieval-augmented generation across very large inputs. GreenPT prices it at roughly $1.25 per million input tokens and $5 per million output tokens, aligning it with mid-range hosted pricing and keeping it competitive for sustained production traffic. The combination of a deep context window, agent-oriented capability flags, and open-weight availability makes GLM-5.2 Caveman Lite a sensible fit for teams that want a flexible GreenPT-hosted model for chat, tooling, and structured-output workloads without paying for the highest output tier.