Models

LLMBase routes your request to the best available upstream provider automatically. You reference models by their unified ID — provider details are never exposed.

Available models

Model ID	Context window	Max output	Best for
zai-org/glm-5	131,072 tokens	16,384 tokens	Complex reasoning, long context, bilingual (ZH/EN)
zai-org/glm-4	131,072 tokens	8,192 tokens	Everyday tasks, fast responses, bilingual (ZH/EN)

Choosing a model

• zai-org/glm-5 — Use for tasks that require deep reasoning, multi-step instructions, or long documents. Higher quality, slightly more expensive.
• zai-org/glm-4 — Use for shorter, high-frequency requests where latency and cost matter more than maximum capability.

List models — GET /v1/models

Returns all available models in the OpenAI models format. Requires authentication.

curl https://api.llmbase.ai/v1/models \
  -H "Authorization: Bearer $LLMBASE_API_KEY"

Response

{
  "object": "list",
  "data": [
    {
      "id": "zai-org/glm-5",
      "object": "model",
      "created": 1700000000,
      "owned_by": "zai-org",
      "name": "Z.ai: GLM-5",
      "description": "High-performance bilingual instruction model with strong reasoning capabilities.",
      "context_length": 131072
    },
    {
      "id": "zai-org/glm-4",
      "object": "model",
      "created": 1700000000,
      "owned_by": "zai-org",
      "name": "Z.ai: GLM-4",
      "description": "Efficient bilingual instruction model with strong reasoning and tool-use capabilities.",
      "context_length": 131072
    }
  ]
}

With the OpenAI SDK

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.llmbase.ai/v1",
  apiKey: process.env.LLMBASE_API_KEY,
});

const models = await client.models.list();
for (const model of models.data) {
  console.log(model.id);
}

Automatic failover

Each model is backed by multiple upstream providers. If the primary provider is unavailable or rate-limited, requests are transparently retried against the next provider. You never need to handle this yourself.