Unified AI API

One unified API for multiple AI models

Write your integration once. LLMFlux exposes every model in its catalogue through the same endpoint, the same key and the same request formats, so moving between models is a one-field change instead of a new integration.

Why one API for many models

Each AI provider ships its own SDK, authentication, error format and billing. Supporting three providers means maintaining three integrations, and trying a new model means writing a fourth. A unified API removes that cost: the request shape is the same for every model, so model choice becomes configuration.

  • Same request, any model

    Every catalogue model accepts Chat Completions, Responses and Messages requests.

  • One credential

    A single LLMFlux key instead of one key per provider.

  • One balance

    Prepaid credits cover every model; each request is priced per model.

  • One log

    All requests, across all models, in one searchable dashboard.

Models available through the unified API

The catalogue currently lists 28 models from 10 model makers (Anthropic, DeepSeek, Google, MiniMax, NVIDIA, Qwen, TeleAI, Upstage, xAI, Zhipu AI). Availability changes as models are added or retired; the list below is read from the live catalogue.

Model idMakerContext window
claude-fable-5Anthropic200,000 tokens
claude-opus-4-8Anthropic200,000 tokens
claude-opus-5Anthropic200,000 tokens
claude-opus-5-5Anthropic200,000 tokens
claude-sonnet-5Anthropic200,000 tokens
claude-sonnet-5-5Anthropic200,000 tokens
deepseek-unrestrictedDeepSeek131,072 tokens
deepseek-v4.1-flashDeepSeek131,072 tokens
deepseek-v4-flashDeepSeek131,072 tokens
gemini-3.8-flashGoogle1,000,000 tokens
minimax-m2.7MiniMax131,072 tokens
nemotron-3.5-lightningNVIDIA131,072 tokens
nemotron-3.5-super-vlNVIDIA131,072 tokens
nemotron-3-superNVIDIA131,072 tokens
nemotron-3-ultra-550b-a55bNVIDIA131,072 tokens
nemotron-3-ultra-unrestrictedNVIDIA131,072 tokens
qwen3.8-27bQwen131,072 tokens
qwen-unrestrictedQwen131,072 tokens
xing4.0-29b-a4bTeleAI131,072 tokens
solar-mini4Upstage131,072 tokens
solar-pro4Upstage131,072 tokens
solar-unrestrictedUpstage131,072 tokens
grok-4.6xAI131,072 tokens
grok-4.7xAI131,072 tokens
grok-unrestrictedxAI131,072 tokens
glm-5.3-flashZhipu AI131,072 tokens
glm-5.3-flash-2Zhipu AI131,072 tokens
glm-unrestrictedZhipu AI131,072 tokens

Prices per 1M tokens are on the model catalogue. With a key, the same list is available from the API:

List modelsbash
curl https://api.llmflux.dev/v1/models \
  -H "Authorization: Bearer $LLMFLUX_API_KEY"

Switching models in code

Because the format is shared, comparing or switching models is a loop over model ids:

Python (openai)python
# Same request, different model: only the "model" field changes.
for model in ["claude-sonnet-5-5", "deepseek-v4-flash", "grok-4.7"]:
    resp = client.chat.completions.create(
        model=model,
        messages=[{"role": "user", "content": "Reply with your model name."}],
    )
    print(model, "→", resp.choices[0].message.content)

In TypeScript the client is created once and reused for every model:

Node.js (openai)ts
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.llmflux.dev/v1",
  apiKey: process.env.LLMFLUX_API_KEY,
});

const resp = await client.chat.completions.create({
  model: "claude-sonnet-5-5",
  messages: [{ role: "user", content: "Hello!" }],
});
console.log(resp.choices[0].message.content);

What developers build with a unified AI API

  • Model evaluation

    Run the same prompts against several models and compare quality, latency and cost from the request log.

  • Tiered workloads

    Send classification and extraction to a fast, low-cost model and reasoning-heavy tasks to a larger one.

  • Model upgrades

    Move to a newer model version by changing configuration, then roll back the same way.

  • Multi-model products

    Let your own users choose a model in your UI without you integrating each provider.

Documentation

Frequently asked questions

What is a unified AI API?
A unified AI API (also called a unified LLM API or multi-model AI API) exposes many AI models through one endpoint, one authentication method and one request format. Your code stays the same while the model changes.
How do I switch AI models using one API?
Send the same request with a different value in the model field. Every model in the LLMFlux catalogue accepts the OpenAI Chat Completions and Responses formats and the Anthropic Messages format, so no other change is needed.
Which models are included?
The live list is at llmflux.dev/models and from GET /v1/models with your key. It currently includes models from Anthropic, DeepSeek, Google, MiniMax, NVIDIA, Qwen, TeleAI, Upstage, xAI, Zhipu AI.
Is there a model that picks the best model for me?
Not in the sense of per-request quality prediction. The model id "auto" maps to one configured catalogue entry. We recommend naming the model explicitly so results are predictable.
Do all models cost the same?
No. Each model has its own input, output and cache prices per 1M tokens, shown in the model catalogue. Usage is charged from one prepaid credit balance.

Start with one key

Create an account, generate an API key, and send your first request in a few minutes.