DecisionModelHub
Workspace/Model catalog

Choose a model for your constraints.

Filter documented access and pricing facts, then compare up to three starting points side by side.

4 models
Hosted API

Clef Flash

Cloudflare

A 9B decision model for typed choices, probability checks, and scoring. Hosted on Workers AI.

choiceprobability (API type: noul)score
$0.09 / 1M input tokens
Context window65,536 tokens

Good starting context: Hosted, lower input-price starting point for typed decisions.

Support routing · 100 cases ·

87/100 correct87% (79–92) accuracyAccuracy of valid answers · 95% interval195 ms median
Measured profileFull benchmark
Explore schema Provider docs
Hosted API

Clef

Cloudflare

A 27B decision model with the same typed question interface. Evaluate it on the decisions your application actually makes.

choiceprobability (API type: noul)score
$0.24 / 1M input tokens
Context window65,536 tokens

Good starting context: Hosted larger-model alternative to evaluate against Clef Flash.

Support routing · 100 cases ·

92/100 correct92% (85–96) accuracyAccuracy of valid answers · 95% interval271 ms median
Measured profileFull benchmark
Explore schema Provider docs
Hosted API

Jev

TypeSafe AI

TypeSafe’s System One model. Turns application state and typed questions into structured decisions.

choiceprobability (API type: noul)score
$0.042 / 1M input tokens
Context windowUnknown (see provider docs)

Good starting context: Hosted TypeSafe API for structured decisions.

Support routing · 100 cases ·

92/100 correct92% (85–96) accuracyAccuracy of valid answers · 95% interval138 ms median
Measured profileFull benchmark
Explore schema Provider docs Pricing source
Hosted API

Decisions API

OpenAI

OpenAI’s decisions endpoint, in public beta and served by GPT-6 Luna. Returns a probability, a choice or a score for text and image input.

choiceprobability (API type: predicate)score
$0.1 / 1M input tokens
Context windowNot stated (GPT-6 Luna: 1,050,000 tokens)

Good starting context: Hosted OpenAI endpoint for typed decisions on text and images.

Support routing · 100 cases ·

94/100 correct94% (88–97) accuracyAccuracy of valid answers · 95% interval111 ms median
Measured profileFull benchmark
Explore schema Provider docs
Provider-reported information · checked October 9, 2026.
These facts narrow a shortlist; they do not name a universal winner. Prices are input-token rates, not total request cost. Jev’s advertised $42 per billion is converted to $0.042 per million. Unknown values are labeled explicitly. Provider terms and availability can change.