Cortex
CortexFoundry · Model tuning

Improve a model on the work you know.

Train a specialized model on accepted and corrected Work Units. Keep the weights in your environment, test it against your acceptance criteria, and route only the work it clears.

Training data

Turn accepted work into training and evaluation data.

When the same Work Units recur, accepted outputs and human corrections can become supervised training examples and a gold evaluation set.

Promote pipeline

Reviewed outcomes

Proven work is ready to become a reusable quality asset.

Support triage answer$420Accepted
People
Agents
Applications
+1
2m ago
Preferred · promote to gold set
Contract clause extraction$1,080Edited
People
Agents
Applications
11m ago
Correction · promote to eval set
Invoice coding batch$640Accepted
People
Agents
Applications
+1
22m ago
Preferred · promote to training corpus
Promoted into reusable assets
Gold setVerified best answersHuman verified
Target modelcortex-triage-v3Last promoted2m agoLast verifiedToday 09:14Label sourceShip events · Preferred only
1,240 rows+38 this week
Eval setEdits & reverts become test casesMixed
Target modelClaude Sonnet 4Last promoted11m agoLast verifiedToday 08:02Label sourceHuman edits + inferred reject paths
380 cases+12 this week
Training corpusAccepted pairs, ready for SFTInferred
Target modelcortex-triage-v3Last promoted22m agoLast verifiedYesterday 18:40Label sourceAuto-labeled Preferred from Atlas
8,900 pairs+310 this week
The model

Compare tuned models against your acceptance criteria.

Evaluate a specialized model on the same task, dataset, and scoring method as your current provider. Promote it only when it clears the bar you set.

Sample model evaluation

1008570$0$0.025$0.05Cost per taskQuality on your task91 · $0.00678 · $0.02484 · $0.02882 · $0.041Your tuned model
Illustrative comparison. Use your workload and criteria.
Serving

Qualified work routes to your model. Everything else falls back.

Matching Work Units route to your private model, with frontier models and humans as fallback.

  • Qualified work routes to your model
  • Frontier models and humans as fallback
  • Only proven work is auto-routed
Cortex
Work unit
Support triage
People
Applications
Cortex router
cortex-triage-v3Owned
Fallback lanes
ChatGPT-4.1(Frontier model)if quality gate fails
Human queueif confidence low

Ready to build sovereign intelligence?

Gain visibility and control. Build better, cheaper, faster AI that you own.