Jev 1.13 on QuickSilver Pro
Jev 1.13 is TypeSafe's System One model. It does not write text: you send a state and a set of typed questions, and it answers each one in a single forward pass — pick one of your options, place the state on your scale, or return a yes-leaning probability — with calibrated probabilities attached. Call it on POST /v1/systemone with the same QuickSilver Pro key and USD balance as the rest of the catalog: $0.042 per million input tokens, and output tokens are free.
At a glance
Routing, classification, triage and guardrail decisions as calibrated probabilities — nothing to parse.
Pricing comparison ($/1M tokens)
| Provider | Input | Output | vs QSP |
|---|---|---|---|
| QuickSilver Pro | $0.042 | Free | — |
| TypeSafe list price (jev-1.13) | $0.042 | Free | same |
When to use
Reach for Jev 1.13 wherever your code needs a decision rather than prose: routing a request to the right model or queue, classifying or triaging tickets, scoring content against a rubric, moderation and guardrail checks, or choosing an agent's next step. Every answer comes back typed — a choice with a probability per option and a confidence, a score, or a probability — so there is no free text to parse and no JSON to repair, and you can threshold on the probability instead of trusting a bare label. Up to 64 questions can share one state in a single call, and because only input is billed, a decision over a 1,000-token state costs about $0.00004.
When to use something else
Jev 1.13 does not generate text — no answers, summaries, code or tool calls — and it cannot be reached through Chat Completions; use a chat model for anything generative. Its window is 32K tokens, so trim or summarise long documents before asking about them. Responses are not streamed and a call typically takes about 2–3 seconds end to end, so for tight loops put several questions into one request instead of making many calls.
Quickstart: POST /v1/systemone (curl)
curl https://api.quicksilverpro.io/v1/systemone \
-H "Authorization: Bearer $QSP_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "jev-1.13",
"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":"jev-1.13",
# "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": "jev-1.13",
"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
A model that makes decisions instead of writing text. You send a state — any JSON, a string or an object — plus a set of typed questions, and Jev 1.13 answers every question in one forward pass with calibrated probabilities. There are three question types: choice (pick one of the options you name, with a probability for each), score (place the state on a scale you describe) and noul (a probability-like number for a yes-leaning question).
POST https://api.quicksilverpro.io/v1/systemone with your QuickSilver Pro key as a Bearer token and a JSON body holding model, state and questions — up to 64 questions per call. It is a plain HTTPS JSON endpoint, so curl, requests or httpx all work. It is not part of the OpenAI Chat Completions surface: a chat call to jev-1.13 is rejected, and streaming is not supported. The /docs/systemone guide has the full schema, all three question types and error codes.
Input tokens only, at $0.042 per million — output tokens are free. Each response's usage block reports input_tokens, output_tokens and the USD cost of that call, deducted from the same prepaid balance as every other model. That is TypeSafe's own list rate; QuickSilver Pro adds no markup.
Answers are keyed by the question IDs you chose. A choice question returns the picked option, a probability for every option and a confidence — for example {"type":"choice","choice":"yes","probabilities":{"no":0.07,"yes":0.93},"confidence":0.87}. A noul question returns a single number, such as {"type":"noul","noul":0.69}. The response also carries an id, the model and a usage block with the call's cost.