Tasks and inputs
Sage 2 is listed for Text classification, Structured Decisions. It accepts text, image and produces decision. Use the publisher’s input format and preprocessing for your chosen deployment.
Levanto Sage 2 is listed as a new version of Levanto's hosted Sage decision model; Levanto's own docs do not yet name a version 2.
The specification score summarizes recorded features and limits. It is not a benchmark result.
Pricing conditions: No public per-model rate. Levanto bills input tokens, extraction/streaming/reasoning output tokens and grounding searches at plan rates; decision answers are free of output charges. Confirm on your Levanto billing page.
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
Produces
Levanto Sage 2 appears in model catalogues as the next release of Sage, Levanto Labs' hosted decision model. We could not confirm it. Levanto's documentation describes Sage as a product without naming any version, and the newest version identified elsewhere is 1.3.0, reported as levanto-sage-v1.3. Until Levanto publishes a version note, treat "Sage 2" as unverified, and treat its listed release date of 2 November 2026 the same way.
What we can describe is the model family itself.
Sage does not write prose. You send content and a question; it returns a structured answer and a probability. For a yes/no question, the answer is yes, no or null. That null is deliberate. When Sage is unsure, it says so, and Levanto's guidance is to route a null to a human rather than retry.
Supported decision kinds:
Extraction and streaming endpoints sit alongside these, and images are accepted inline as PNG, JPEG or WebP in beta.
A whole request must fit your plan's context, about 32K tokens or 128K with extended context. Oversized requests fail rather than being truncated. Reasoning, when enabled, has a 6-second limit per answer.
Levanto charges per input token, per output token and per grounding search, at your plan's rates. A decision's answer carries no output charge; extraction output, streaming output and reasoning tokens do. No per-model public rate was published, and a third-party price attached to "Sage 2" could not be verified. Check your Levanto billing page.
Ask Levanto which version your key actually serves. Responses may report a version string, but the selectable ID has been plain levanto-sage, not a pinned version. Then run your own labelled set through it and look at the null rate as closely as the accuracy. A model that abstains on 15% of cases is useful only if you have someone to catch them. For context on how this category works, see structured decision models and compare with TypeSafe's Jev.
Compare the model on the work you need to complete.
Sage 2 is listed for Text classification, Structured Decisions. It accepts text, image and produces decision. Use the publisher’s input format and preprocessing for your chosen deployment.
Test labelled examples and choose thresholds from the cost of each error. Validate probabilities against your application’s outcomes.
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.
These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.
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.
Perplexity
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.
Perplexity
pplx-decider-v1.3 is Perplexity's 27B-class decision model: it returns probabilities for choice, yes/no, score and new span-locate questions, not text.
Butler
Cicero is Butler Compute's open-weight MoE decision model: up to 64 typed questions per state, a calibrated probability for every option, no text.
Morph
Morph Reflex is a family of fast classifier models that label agent turns (jailbreak, looping, frustration and more) with per-class scores.
Sarvam
Sarvam Prism is a 105B decision model that returns calibrated probabilities for yes/no, choice and score questions in English and 10 Indian languages.