Cheap inference API.

Visible token pricing, no contracts, no minimums. OpenAI-compatible.

Open-weight models on shared warm endpoints, priced per token. See live pricing below.

Model pricing

Pay per token. No commitments.

Competitive speed vs. OpenRouter-listed providers, at the lowest price in the benchmark snapshot.

Model Context Input Cache Output
MiniMax M2.7 Live now FP8 200K $0.30/M
Kimi K2.6 Live now INT4 262K $0.70/M
GLM 5.1 Live now FP8 203K $0.90/M
Gemma 4 (31B) Live now BF16 262K $0.11/M

More models are coming soon and will be added as they go live.

Integration

One base URL change.

Keep the OpenAI SDK and point it at Lilac. Your existing code just works.

from openai import OpenAI

client = OpenAI(
    base_url="<your_base_url>",
    api_key="sk_...",
)

response = client.chat.completions.create(
    model="openai/gpt-5.4",
    messages=[{"role": "user", "content": "Hello!"}],
)

Same code. Same SDK. Fraction of the price.

  1. OpenAI-compatible — switching is a base URL change.
  2. Shared endpoints stay warm. No cold starts.
  3. No contracts or minimums. Start immediately.

Frequently asked questions

What makes Lilac cheap?

We route inference to idle enterprise GPUs — hardware already powered on and paid for.

Does cheap mean slower?

No. We benchmark competitively with OpenRouter-listed providers at the same price point or lower.

Start running inference in minutes.

No contracts, no commitments. Swap your base URL and pay less for the same output quality.