Prefill Systems
Models

Models & pricing

Starting with Qwen3.8-27B. OpenAI-compatible inference with straightforward per-token pricing.

Serving text in / text out streaming + non-streaming

Model card

Public model IDqwen/qwen3.8-27b
APIOpenAI-compatible chat completions — POST {base}/chat/completions (base = https://prefillsystems.com/v1)
Usage reportingPrompt, cached-prompt, and completion counts reported by the inference worker
AuthenticationBearer API key (invitation only; see docs)
AccessInvitation-only pilot · OpenRouter distribution planned
Quantization & contextEngine-dependent; we do not publish values we have not locked for production hosting

Full model metadata is served to authenticated clients at GET {base}/models.

Pricing

Rates are per one million tokens (USD) and match the live gateway accounting configuration exactly:

Token typeUSD per 1M tokensNotes
Input — uncached prompt $0.24 Applied to prompt tokens not reported as cached by the engine
Input — cached prompt $0.24 Applied to engine-reported cached prompt tokens when the backend reports them
Output — completion $2.15 Applied to generated completion tokens

List models

Authenticate with a dashboard-generated key and list what the gateway serves:

# replace YOUR_API_KEY with a key from your dashboard
curl https://prefillsystems.com/v1/models \
  -H "Authorization: Bearer YOUR_API_KEY"

Response:

{
  "object": "list",
  "data": [
    {
      "id": "qwen/qwen3.8-27b",
      "object": "model",
      "created": 0,
      "owned_by": "prefill-systems",
      "pricing": {
        "prompt_per_million": "0.24",
        "cached_prompt_per_million": "0.24",
        "completion_per_million": "2.15"
      }
    }
  ]
}

Honest limits