Skip to main content

Low-latency, general purpose inference.

Celeris-1 is the first model from Celeris: a general purpose language model that delivers near-GPT-5 level intelligence with up to 15x faster response times.

Get startedAPI reference

Make your first call

Use the familiar OpenAI request shape. See the quickstart for SDK setup.

curl https://inference.celeris.ai/celeris-1/v1/chat/completions \
  -H "Authorization: Bearer $CELERIS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "celeris-1",
       "messages": [{"role": "user", "content": "Reply with one word: fast or slow?"}],
       "max_tokens": 256, "temperature": 0, "seed": 7}'

From first request to production

Find the shortest path to the task in front of you.

OpenAI-compatible

Use the chat-completions request format with official OpenAI SDKs and other clients that support a custom base URL.

Making requests

Built for agents

Use celeris-1 for short intermediate tasks such as classification, extraction, scoring, and query rewriting.

Agentic workloads

High-throughput generation

Parallel token generation delivers high tokens per second for real-time code assistance, live translation, voice agents, and bulk document transformation.

Open the cookbook

Latency you can audit

Successful responses include Server-Timing metrics for server-side latency analysis.

Measuring latency