Cerebras can dramatically accelerate generation on supported models, but agent latency still depends on prefill, tool I/O, model quality, and platform compatibility.
For side projects, toy demos, and RAG prototypes, nobody wants to swipe a credit card on day one. This is a verified roundup of 40+ LLM inference providers still operating as of 2026/05, tiered by whether free resources auto-replenish or are one-time grants. Each entry notes credit-card requirements, supported models, paid starting prices, and catches. Chinese-origin providers including Zhipu GLM (permanently free), Doubao (2M tokens/day), Kimi, DashScope, and the Ollama local option are all included.