Juneo LabAn AI safety lab

ForecastIssue 01November 2026

The trust layer forpeople and agents

Laplace foresees the consequence before any agent acts

Consequence forecast

A retention agent wants to give a 15% renewal discount

Annual contract of ¥640,000, cancellation notice received

Organization AAgents may grant up to 20% on their own

  1. Go ahead (Go ahead)97%The discount applies now and the customer renews today
  2. Hand it to a person1.7%The sales director approves first; the customer waits a day
  3. Stop it1.6%No discount; the customer may leave
Laplace decidesGo ahead
Consequence forecastRecorded Laplace-1-Vector output2026-11-110.05 s

From inside one organizationto any two agents

Open a scenario to see how Laplace decides at each step

Go ahead
Go ahead with changes
Ask a person first
Stop

Illustrative scenarios. People, companies, and data are fictional

Talk to us about your scenario

Organizations delegatemore than actions

Four decisions, each with a recorded forecast

  1. Allocate work

    A large disputed invoice: who takes it?

    1. Sun Hui, receivables (Go ahead)98%
    2. Billing agent1.8%
    3. Research agent<1%
  2. Share information

    A supplier's agent asks for costs, stock and forecasts

    1. The 90-day forecast only (Go ahead with changes)99%
    2. Nothing1.3%
    3. Everything<1%
  3. Act

    A procurement sub-agent tries to transfer ¥38,000

    1. Route to the finance agent (Ask a person first)77%
    2. Stop it21%
    3. Let it transfer1.7%
  4. Evaluate outcomes

    An agent reports 40 refunds done; there are 37 receipts

    1. Send back for 3 receipts (Go ahead with changes)98%
    2. Reject and reverse1.8%
    3. Accept and close<1%

People and companies in the inputs are fictional; the probabilities are recorded Laplace-1-Vector outputs

See how Laplace works

First on mostdecision benchmarks

A flagship and an on-device model, with complete open weights

Laplace-1-Atlas and models of its class: the best score on each decision benchmark
Laplace-1-AtlasJuneoDecider v1.2PerplexityClef-OmniCloudflareJev 1.14TypeSafeGPT-6 Luna DecisionsOpenAIDoubao Seed 2.1 LiteByteDancedecision-model-previewAlibaba
JevBench93.2Best, ties included93.2Best, ties included86.9Not best91.6Not best89.5Not best92.7Not best77.0Not best
JEVal79.2Not best82.2Best, ties included77.6Not best80.0Not best—Not published78.3Not best60.2Not best
When2Call89.9Best, ties included87.6Not best62.4Not best81.3Not best74.5Not best70.7Not best45.2Not best
AgentDecisionBench L2 common95.6Best, ties included94.1Not best48.4Not best91.0Not best90.9Not best85.8Not best43.0Not best
AgentDecisionBench L2 shift95.5Best, ties included94.0Not best52.3Not best92.1Not best88.2Not best87.8Not best48.5Not best
Coherence85.7Best, ties included80.8Not best—Not published82.7Not best—Not published80.6Not best—Not published
COMPASS85.3Best, ties included84.3Not best81.8Not best81.3Not best—Not published79.1Not best73.7Not best
OrgAccess44.8Not best30.9Not best35.7Not best42.0Not best45.7Best, ties included41.2Not best16.1Not best
TrustDecision89.4Best, ties included84.0Not best78.8Not best78.7Not best—Not published79.6Not best61.8Not best
  1. Laplace-1-AtlasJuneo
  2. Decider v1.2Perplexity
  3. Clef-OmniCloudflare
  4. Jev 1.14TypeSafe
  5. GPT-6 Luna DecisionsOpenAI
  6. Doubao Seed 2.1 LiteByteDance
  7. decision-model-previewAlibaba

Benchmark order: 1 JevBench,2 JEVal,3 When2Call,4 AgentDecisionBench L2 common,5 AgentDecisionBench L2 shift,6 Coherence,7 COMPASS,8 OrgAccess,9 TrustDecision

Best, ties includedNot bestSource: Laplace-1 model cards

ModelSizeLedDownloads
Laplace-1-Atlas27B7 / 9—
Laplace-1-Vector4B7 / 8—

The standard of judgmentlives inside the organization

Decision Forge turns an organization's own policies, approvals and tickets into verified decision training data. Laplace-1 is trained entirely on it

The organization's material

  • Policies
  • Approvals
  • Tickets
  • Conversations
  • Logs

Decision Forge

  1. 1Parse
  2. 2Extract decision events
  3. 3Model rules and authority
  4. 4Counterfactual expansion
  5. 5Verify by policy replay

Verified training data

Decision context
State and permissions when an agent acts
Judgment standard
What is allowed, forbidden, or escalated
Execution receipt
What actually changed afterwards

Trains Laplace-1

Judgments organizationshave already handed to AI

With front-line teams in manufacturing, education, healthcare and home renovation, we decide which judgments can go to AI and how the trust layer checks them

Let your organizationtrust agents with real work

Delivered for 18 customers in education, tax, manufacturing, healthcare and home renovation

Contact sales