SSyncDev

Solar Decide

SyncDev model directory · Profile updated Nov 18, 2026

Compare with another model

Solar Decide is Upstage's beta decision model on Solar Mini 4: typed questions in, a calibrated choice, score or yes/no out, 512K context, free output.

Key specifications

Specification score
28.1
Context window
512K
Input / 1M tokens
From $0.10
License
Proprietary

The specification score summarizes recorded features and limits. It is not a benchmark result.

Pricing conditions: Indicative: $0.10 per 1M input tokens, output free (beta). One aggregator lists $0.05; confirm on your Upstage account.

Inputs and outputs

Recorded modality support shows what this model can accept and produce. An undocumented modality is shown as unknown rather than assumed to be unsupported.

Accepts

  • Text input

Produces

  • Text output

What is Solar Decide?

Most classification pipelines ask a chat model to write the answer, then parse the text and hope it chose a valid label. Solar Decide removes the writing step entirely. You send a state and a set of typed questions; it returns a choice, a score or a yes/no, each with a probability read directly from the model. No prose is generated at all.

Upstage launched it in beta on 22 September 2026, the same day as its compact Solar Mini 4. The decision endpoint runs on that model as a "System One" service.

How a request is shaped

The API uses the POST /v1/systemone schema, which Upstage says is the same schema as TypeSafe's Jev. That is a practical detail: if you have already integrated a Jev-style decision call, the migration is mostly a base-URL and model-name change.

Field Value
API ID solar-decide
Status Beta, Upstage Console only
Released 22 September 2026
Context 512K tokens
Question types Choice, score, yes/no
Input $0.10 per 1M tokens (cached also $0.10)
Output Free

The half-million-token window is the selling point. A full contract, a long support thread or a policy manual can be the state, and you can ask several typed questions against it in one call.

Where the numbers are thin

Upstage calls the probabilities calibrated but does not explain the method or publish a calibration benchmark. Nobody outside the company has measured it either, as far as public material goes. One aggregator lists the input price at $0.05 per million, half of what the Upstage console shows; the console figure is the one to plan with, and both should be confirmed against your own bill.

Language coverage is also vague. The model page stresses strong Korean understanding and stops there.

A test plan that earns trust

  1. Take a labelled set of 300 to 500 past decisions from your own process.
  2. Send each as a state with the same typed question and record the returned probability.
  3. Bucket the predictions by probability and check that items scored 0.8 are right about 80% of the time.
  4. Shuffle answer options on a subset and confirm the top choice does not move.

If the reliability curve holds, you can route on thresholds: auto-approve high-confidence items, send the middle band to a person. If it does not, the free output is not much of a bargain.

For other decision-style endpoints, browse the structured decisions category. Upstage's general-purpose flagship is Solar Pro 4, and the rest of the line-up sits on the Upstage provider page. Treat all pricing here as indicative, since beta terms can change.

How to evaluate Solar Decide

Compare the model on the work you need to complete.

Tasks and inputs

Solar Decide is listed for Text classification, Structured Decisions. It accepts text and produces text. Use the publisher’s input format and preprocessing for your chosen deployment.

Capacity and cost

Documented limits: 512K context. The listed route has starting rates of $0.10 input per million tokens. Measure a completed task using the prompts, documents and tools your application will use. Include reasoning and retries in the result.

Choosing a deployment

Use the model identifier and account requirements for the selected API route. Confirm its request limits and output format with one complete task before connecting a larger workflow.

Frequently asked questions

A choice, a score or a yes/no answer for each typed question, each with a calibrated probability. It generates no text.

These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.

Microsoft

Microsoft-Decision-1.1

Microsoft-Decision-1.1 is a decision model for routing, triage and classification, returning calibrated probabilities. GA in Foundry from November 2026.

— context

Liquid AI

d1.2

Liquid AI d1.2 is a hosted omnimodal decision model for text, image and audio questions with calibrated answers, at $0.02 per million input tokens.

65.54K context

Liquid AI

d1.2-ultra

Liquid AI d1.2-ultra is the top hosted decision model in Liquid's line, scoring 75 on the AA Decision Index, with audio support at $0.042 per 1M input.

65.54K context

Google

Gemini Decision 1

Gemini Decision 1 powers Google's Decisions API, returning calibrated yes/no, choice and rubric answers instead of text, at $0.10 per 1M input tokens.

262.14K context

subquadratic

SubQ Snap 1.2

SubQ Snap 1.2 is Subquadratic's long-context decision model on linear sparse attention, tested to 256K tokens, priced at $0.04 per 1M input tokens.

— context

Perplexity

pplx-decider-flash-v1.3

pplx-decider-flash-v1.3 is Perplexity's small, half-price decision model, distilled from the 27B decider, for high-volume classification and span extraction.

— context