Skip to navigation
Authors and ModelsEmpirioLabs

Aplomb 1

Aplomb 1
EmpirioLabs · Decision Models
POST /v1/decisions

Answers yes/no, choice, score and tool questions about text, JSON, images, video and audio in up to 1M tokens, and can tell when the input lacks the answer.

At a glance

FieldValue
Model idaplomb-1
Model release date2026-09-29
Input modalitiesText, Image, Video, Audio
Output modalitiesDecision
Context window1M
Weight precisionBF16 weights
Featuresclassification, calibrated_probabilities, tool_selection, abstain, long_context, vision, video, audio_in, multimodal
Native inferenceYes
NewYes
Supported endpointsPOST /v1/decisions, POST /v1/chat/completions, POST /v1/responses, POST /v1/messages, POST /v1beta/models/aplomb-1:generateContent
Alternate model idsaplomb, empiriolabs/aplomb-1, empiriolabs/aplomb

Pricing

ChargeSpecRate
Inputper 1M prompt tokens$0.02
Outputper 1M generated tokensFree

Example request

curl https://api.empiriolabs.ai/v1/decisions \
  -H 'Authorization: Bearer $EMPIRIOLABS_API_KEY' \
  -H 'Content-Type: application/json' \
  -d '{"model": "aplomb-1", "state": "Customer: I was charged twice for my subscription this month and I need a refund today.", "questions": {"is_urgent": {"type": "noul", "instructions": "Does the customer need a resolution today?"}, "department": {"type": "choice", "instructions": "Which team should handle this message?", "criteria": {"billing": null, "technical": null, "sales": null}}}}'

Parameters

ParameterTypeRequiredDefaultDescription
long_contextenumno"fast"How a state longer than 131,072 tokens is read: fast (the default) answers in seconds, full reads every token. Shorter states are always read in full. · Allowed: fast, full

Notes

Requests

Send content in state (a string, a JSON object, or an array) and 1 to 128 questions in questions to POST /v1/decisions. Each question is answered on its own, with probabilities and a confidence value. No text is generated. The Decisions API guide documents every question option, and its Chat formats section covers requests through /v1/chat/completions, /v1/responses, /v1/messages and Gemini generateContent.

Question types

  • noul: yes or no, answered with the probability of yes.
  • choice: one of 2 to 255 options, answered with the chosen option and the full distribution.
  • score: 2 to 10 ordered levels, answered with the expected level and the distribution.
  • tool: which of 1 to 254 function tools to call, with distributions over its enum and boolean arguments.

Not in the state

Add "abstain": true to a noul, choice or score question to let the model say the state does not contain the answer. The answer then includes abstain, the probability of that outcome, and the rest of the answer assumes the state does contain it.

Limits

  • The state plus the longest question: up to 1,000,000 tokens.
  • At most 512 evaluated questions per request. A tool question counts as one plus one per enum or boolean argument.
  • Media: up to 16 images (20 MB each), 8 audio clips (25 MB and 10 minutes each) and 4 videos (100 MB and 10 minutes each), as public HTTPS URLs or inline base64.

Benchmarks

Benchmark results are in the launch post.

Billing

Input tokens only: the state once per request, each question’s own text, and each media item’s tokens. Output is always zero.

Machine-readable schema: GET https://api.empiriolabs.ai/v1/models/aplomb-1.