Lumino Learn

Practical GPU and AI infrastructure guides

Use these guides to plan VRAM, model serving, hosted APIs, GPU rental, and inference costs before your product traffic arrives.

GLM API guide

GLM 5.2 API Guide: Long Context, Coding Agents, Pricing and When to Use It

A practical GLM 5.2 API guide for developers comparing long-context coding agents, OpenAI-compatible access, cached-token pricing, local serving, and GPU rental tradeoffs.

Read guide

LLM infrastructure guide

VRAM Calculation for LLMs: Weights, KV Cache, Context and Batch Size

A practical guide to estimate GPU VRAM for LLM inference before renting hardware, including model weights, quantization, KV cache, context length, and runtime overhead.

Read guide

Long-context inference

Long Context Is a GPU Memory Problem Before It Is a Prompting Feature

Why long-context AI apps hit GPU memory, TTFT, and throughput limits, and how to plan KV cache, batching, and fallback APIs before users arrive.

Read guide

Inference cost control

LLM Inference Cost Leaks: The GPU Is Not Always the Expensive Part

A practical guide to the hidden cost leaks in LLM inference: idle GPUs, long outputs, retries, cold starts, oversized context, and unmanaged routing.

Read guide