Origin Docs
Console

Model List

Browse the ORGN Gateway model catalog in the console. Filter by provider, execution type (TEE or ZDR), and modality. Model IDs use underscores.

The Model List in the Gateway console is the authoritative browser for every model available through ORGN Gateway. Use it to discover models, compare execution types, and copy model IDs for API requests.

The same catalog is available programmatically via GET https://api.gateway.orgn.com/v1/models with your sk-ollm-* API key.

Model ID format

All Gateway model IDs use underscores, not slashes:

near_qwen3_30b
phala_deepseek_r1
vercel_claude_sonnet_4_6
vercel_gpt_5

Use the ID exactly as displayed when calling the API or configuring SDK integrations. Slashes and provider-prefixed paths (for example near/gpt-oss-120b) are not valid Gateway identifiers.

Execution types

Not every model runs in a Trusted Execution Environment. The Model List tags each entry with its execution type:

TEE (Trusted Execution Environment)

TEE models run on NEAR and Phala infrastructure inside hardware-isolated Trust Domains. Every request produces a cryptographic attestation receipt you can verify independently in scanner.orgn.com.

PropertyDetail
Prefixesnear_*, phala_*
PrivacyHardware-enforced — encrypted in memory, invisible to host and operators
AttestationYes, per request
VerificationScanner — not available in the Gateway console
Best forRegulated workloads, audit requirements, sensitive data

ZDR (Zero Data Retention)

ZDR models run on Vercel's AI infrastructure under zero data retention provider agreements. Prompts and outputs are not stored or logged, but there is no hardware attestation receipt.

PropertyDetail
Prefixesvercel_*
PrivacyPolicy-enforced — contractual zero retention with model providers
AttestationNo
VerificationNot applicable
Best forFrontier closed-weight models, multimodal, image, and embedding workloads

Choosing a ZDR model means you rely on provider agreements, not cryptographic proof. For workloads that require independent verification, select a TEE model and confirm attestation in Scanner.

Browsing the catalog

In the console Model List you can:

  • Search by model name or ID
  • Filter by provider (NEAR, Phala, Vercel)
  • Filter by execution type (TEE or ZDR)
  • Filter by modality (text, image, audio, embedding)
  • Inspect per-model metadata: context window, output limits, cost per token, and capability tags

Each model entry shows:

FieldDescription
IDAPI identifier (underscore format)
Display nameHuman-readable model name
Providernear, phala, or vercel
Execution typeTEE or ZDR badge
ModalitiesInput and output types (text, image, audio, embedding)
Token limitsMax input and output tokens
CostPer-token pricing where applicable

Programmatic discovery

Fetch the live catalog from your application:

curl https://api.gateway.orgn.com/v1/models \
  -H "Authorization: Bearer sk-ollm-YOUR_API_KEY"
list-models.js
const res = await fetch("https://api.gateway.orgn.com/v1/models", {
  headers: { Authorization: "Bearer sk-ollm-YOUR_API_KEY" },
});
const { data } = await res.json();

for (const model of data) {
  console.log(model.id, model.owned_by);
}

The catalog changes as new providers and model versions are added. Prefer runtime discovery over hardcoded IDs in production applications.

Choosing a model

RequirementRecommended typeExample IDs
Cryptographic proof of executionTEEphala_deepseek_r1, near_qwen3_30b
Frontier closed-weight modelsZDRvercel_claude_sonnet_4_6, vercel_gpt_5
Image or multimodal inputZDR (most)vercel_claude_sonnet_4_6
EmbeddingsBothnear_qwen3_embedding_0_6b (TEE), vercel_text_embedding_3_small (ZDR)
Speech-to-textTEEnear_whisper_large_v3

See Model catalog for modality-specific pages and a detailed TEE vs ZDR comparison.

Testing before integration

  1. Find a model in the Model List and note its ID and execution type.
  2. Open the Playground and send a test prompt.
  3. Create an API key and integrate the model into your application.
  4. For TEE models, verify attestation in scanner.orgn.com.

What's next

On this page