Trio-Spark v1.3 on QuickSilver Pro
Trio-Spark v1.3 is MachineFi's System One decision model. Send text as a state with typed questions to POST /v1/systemone: one forward pass returns choice, score or noul answers with calibrated probabilities, without generating tokens. QSP supports text input only, within a 1,024-token input limit. Pay $0.042 per million input tokens; output is free.
At a glance
Routing, classification and scoring over short text with calibrated probabilities.
Pricing comparison ($/1M tokens)
| Provider | Input | Output | vs QSP |
|---|---|---|---|
| QuickSilver Pro | $0.042 | Free | — |
| MachineFi list price (trio-spark-v1.3) | $0.042 | Free | same |
When to use
Use Trio-Spark v1.3 for routing, ticket triage, classification and rubric-based scoring when the input fits 1,024 tokens. MachineFi reports 90.9% on a ten-task panel of 2,977 decisions, up from 83.1% for v1.0, with expected calibration error 0.019. The same panel reports Jev 1.13 at 90.0%, Clef at 90.8% and Clef-Flash at 91.8%. Group related questions about one state in a single request.
When to use something else
The 1,024-token input limit includes the request's state and questions; requests over it are rejected. Jev 1.13 offers 32K context for longer decision inputs. QSP accepts text input only for Trio: image and video input are unavailable here. Use a chat model for generated text, code or summaries. System One responses do not stream.
Quickstart: POST /v1/systemone (curl)
curl https://api.quicksilverpro.io/v1/systemone \
-H "Authorization: Bearer $QSP_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "trio-spark-v1.3",
"state": "hi",
"questions": {
"greeting": {
"type": "choice",
"instructions": "Is the state a greeting?",
"criteria": {"yes": "it is a greeting", "no": "it is not a greeting"}
}
}
}'
# {"id":"so-45a870bb532d4a72b41404cc8911c72e","model":"trio-spark-v1.3",
# "answers":{"greeting":{"type":"choice","choice":"yes",
# "probabilities":{"no":0.07,"yes":0.93},"confidence":0.87}},
# "usage":{"input_tokens":314,"output_tokens":33,"cost":...,"currency":"USD"}}import os, requests
resp = requests.post(
"https://api.quicksilverpro.io/v1/systemone",
headers={"Authorization": f"Bearer {os.environ['QSP_KEY']}"},
json={
"model": "trio-spark-v1.3",
"state": {"ticket": "I was charged twice for my subscription this month."},
"questions": {
"queue": {
"type": "choice",
"instructions": "Which team should handle this ticket?",
"criteria": {
"billing": "payments, charges, refunds, invoices",
"technical": "bugs, errors, integrations",
"account": "login, email, profile changes",
},
},
"urgent": {
"type": "noul",
"instructions": "Does this ticket need a reply within the hour?",
},
},
},
timeout=30,
)
resp.raise_for_status()
data = resp.json()
queue = data["answers"]["queue"]
print(queue["choice"], queue["probabilities"]) # picked queue + per-option probabilities
print(data["answers"]["urgent"]["noul"]) # probability-like number
print(data["usage"]) # input_tokens, output_tokens, costSystem One API, not Chat Completions: send a state and typed questions, get typed answers with probabilities. Same key and balance as every other model. Full System One docs →
FAQ
Trio-Spark v1.3 returns typed decisions instead of generated text: choice selects from your options with probabilities, score evaluates the state against a list of criteria, and noul returns a yes-leaning probability. Multiple questions can share one state in a single request.
POST https://api.quicksilverpro.io/v1/systemone with your QuickSilver Pro key as a Bearer token and JSON containing model="trio-spark-v1.3", state and questions. Send up to 64 questions within the 1,024-token input limit. QSP supports text input only. See /docs/systemone for the schema. Chat Completions and streaming are unsupported.
Input tokens cost $0.042 per million and output is free, at MachineFi list price with no markup. The response usage block reports input_tokens, output_tokens and the USD cost, charged to your existing QSP balance.
Answers are keyed by your question IDs. A choice answer includes the selected option and probabilities, for example {"type":"choice","choice":"yes","probabilities":{"no":0.07,"yes":0.93}}. A noul answer contains a value such as {"type":"noul","noul":0.69}. The QSP response also includes an id, model and usage block with the cost of the call.