Attestation-backed inference for regulated workloads

Secure & Confidential AI inference
you can verify

Access open-source and frontier-class AI workloads through one secure platform. Enclave42 helps startups and enterprises deploy AI with wallet-based usage control, model-tier access, and real-time monitoring. Every model on Enclave42 runs inside an Intel TDX enclave on NVIDIA GPUs. Generate your own nonce, download the attestation evidence, and verify independently that your prompts were unreadable — including by us.

inference.ts200 OK · 480ms
// One API. Any model tier.
const res = await enclave.chat({
  model: "enclave-frontier-x",
  wallet: "wlt_9f2a…",// PAYG
  messages: [{ role: "user",
    content: "Summarize Q3 filings" }],
  confidential: true// TEE
});
Wallet balance
$1,284.20
2.4M tokens this month
Latency (p95)
API-first developer workflowWallet-based PAYG controlTEE-ready confidential compute
Platform

From model APIs to confidential workloads

Start with secure serverless model APIs. Scale into advanced open-source models, selected TEE-enabled inference, and custom GPU-backed workloads when your use case requires it.

Secure Model APIs

Run language, reasoning, code, vision, embeddings, speech, reranking, and multimodal workloads through one unified API.

Frontier-Class Open-Source Models

Access advanced model families built for production AI workflows, with plan-based access and clear usage controls.

Confidential Compute

Selected TEE-enabled model access and isolated compute options for sensitive, regulated workloads that need stronger guarantees.

Custom GPU Workloads

Business, Sovereign, and Enterprise customers can request private deployments, dedicated GPUs, and custom model hosting through a reviewed engagement.

UAE / GCC Support

Built for AI teams in the UAE and GCC

Enclave42 combines secure AI infrastructure with local consulting, onboarding, and integration support for startups, enterprises, and regulated organizations across the GCC.

  • UAE/GCC enterprise onboarding
  • AI integration and architecture support
  • Security and workload review
  • Cost optimization guidance
  • Private deployment planning for enterprise customers

Regional infrastructure support

Dedicated regional infrastructure tailored for the compliance, performance, and security requirements of ambitious AI teams.

How it works

Subscription unlocks access. Wallet controls usage.

Your subscription unlocks platform features, model tiers, limits, and support. Wallet/PAYG controls real API consumption.

1

Create your workspace

Set up an isolated environment for your team.

2

Choose your plan

Unlock model tiers, limits, and support.

3

Generate an API key

Securely authenticate your applications.

4

Select a model

Use the technical Model ID in your API calls.

5

Top up your wallet

Add funds for pay-as-you-go inference consumption.

6

Monitor usage and cost

Track requests, tokens, cost, and wallet balance.

Security First

Confidential AI for sensitive workloads

For eligible workloads, Enclave42 supports selected TEE-enabled model access and confidential compute options designed for stronger workload isolation, cryptographic security controls, and enterprise-grade monitoring.

Verify our attestation evidence yourself
Availability depends on plan, model type, workload requirements, and enterprise review.
Model Catalog

Available models and price ranges

The same catalog exposed through the developer API. Prices are indicative ranges per 1M tokens; usage is billed from your wallet.

Synced with API

Price range

Filter models by maximum input and output price per 1M tokens.

Input price (max)$3.90 / 1M
$0.00$3.90
Output price (max)$19.50 / 1M
$0.00$19.50
ModelCategoryContextPrice / 1MTierStatus
Mistral Nemo TEETEE
mistral-nemo-tee
Entry TEE131K$0.03 / $0.13
in / out
EntryAvailable
Qwen3 32B TEETEE
qwen3-32b-tee
Standard TEE41K$0.14 / $0.54
in / out
StandardAvailable
Gemma 4 31B TEETEE
gemma-4-31b-tee
Standard TEE131K$0.16 / $0.48
in / out
StandardAvailable
DeepSeek V3.2 TEETEE
deepseek-v3-2-tee
Standard TEE131K$1.30 / $1.30
in / out
StandardAvailable
GLM 5.1 TEETEE
glm-5-1-tee
Advanced TEE203K$1.27 / $4.00
in / out
AdvancedAvailable
Qwen3.6 27B TEETEE
qwen3-6-27b-tee
Advanced TEE262K$0.39 / $2.60
in / out
AdvancedAvailable
Qwen3.5 397B TEETEE
qwen3-5-397b-tee
Selected Frontier TEE262K$0.58 / $3.90
in / out
FrontierAvailable
Kimi K2.6 TEETEE
kimi-k2-6-tee
Selected Frontier TEE262K$0.86 / $4.55
in / out
FrontierAvailable
Nemotron-3-Nano-Omni-30B-TEETEE
nemotron-3-nano-omni-30b-tee
general—$0.03 / $0.13
in / out
EntryAvailable
DeepSeek-V4-Flash-0731-TEETEE
deepseek-v4-flash-0731-tee
general1049K$0.18 / $0.36
in / out
StandardAvailable
GLM-5.2-TEETEE
glm-5-2-tee
reasoning1049K$1.63 / $5.13
in / out
FrontierAvailable
moonshotai/Kimi-K3-TEETEE
kimi-k3-tee
reasoning1049K$3.90 / $19.50
in / out
FrontierAvailable
Qwen3-235B-A22B-Thinking-2507-TEETEE
qwen3-235b-a22b-thinking-2507-tee
reasoning—$0.39 / $1.55
in / out
FrontierAvailable
Showing 13 of 13 modelsPricing is indicative and may vary by workload, region, and plan.

Wondering which model is fastest? We measured time to first token, total latency and throughput across three models and three prompt lengths — 270 requests, published with the raw data and the script to reproduce them.

See the measured latency figures

Beyond model APIs: custom GPU workloads

For Business, Sovereign, and Enterprise customers, Enclave42 supports private dedicated inference endpoints, custom model hosting, long-running jobs, and GPU-backed enterprise workloads through a reviewed engagement.

Available through enterprise review
Self-Service Deployment

Playground

Deploy your own LLM as a private, confidential-compute endpoint. Point to a model, size the GPU, set scaling and access — Enclave42 builds the image and returns a live API URL.

1Model source — Hugging Face repo or custom image
2Compute — GPU type and count, priced by hardware
3Scaling & access — concurrency, auto-scale, visibility
Open Playground
Deployment summary
Deploymentllama-3.1-8b-tee
TemplatevLLM
GPUA100 40GB × 1
VisibilityPrivate
Est. hourly rate$1.10/hr
ONE-TIME DEPLOY FEE$3.30
Plans

Choose platform access, then pay only for real inference usage

Subscription plans define access, model tiers, limits, and support. Wallet balance is used separately for inference usage.

FREE

Explore

$0/mo

Verify our claims before you spend anything.

Includes

  • 5M tokens on signup, valid 30 days
  • 1M tokens every month, ongoing
  • Entry-tier secure models
  • Full attestation verification
  • Unlimited API keys
  • Team workspace — invite colleagues, one shared quota
  • Community support

Not included

  • Standard, Advanced and Frontier models
  • Model deployments
  • Wallet credit

*Platform subscription gives access to Enclave42 features and model tiers. Inference usage is billed separately through your prepaid wallet.

PLATFORM ACCESS

Build

$99/mo

Confidential inference for production apps, at predictable cost.

Includes

  • $25 of inference credit every month
  • Entry, Standard and Advanced secure models
  • Unlimited API keys
  • Usage, wallet and model cost analytics
  • Standard support

Not included

  • Frontier models
  • Model deployments and GPU selection
  • SSO, audit log, DPA

*Platform subscription gives access to Enclave42 features and model tiers. Inference usage is billed separately through your prepaid wallet.

MOST POPULAR

Business

$399/mo

Deploy your own model on the GPU you choose, in an enclave.

Includes

  • $100 of inference credit every month
  • Deploy your own model — Hugging Face repository or custom image, vLLM, SGLang or TEI
  • Choose your GPU — from RTX 3090 to H100 80GB, billed per active compute hour
  • Dedicated private endpoints, with status, technical logs and start/stop control
  • The full catalogue — Entry, Standard, Advanced and selected Frontier models
  • Standard Data Processing Agreement
  • Exportable audit log
  • SSO
  • Priority support

*Platform subscription gives access to Enclave42 features and model tiers. Inference usage is billed separately through your prepaid wallet.

UAE RESIDENT

Sovereign

From $1,499/mo + usage

Provisioned per customer · 30–45 days from signature

For regulated organisations that must prove where their data was processed.

Includes

  • Inference executed on UAE-resident infrastructure, guaranteed in your contract
  • Falcon-H1 Arabic and Jais 2, provisioned on request under the Sovereign engagement, not available in the self-service catalogue
  • Inference credit sized to your contracted capacity
  • Signed, timestamped attestation reports you can archive and hand to an auditor
  • Compliance pack: DPA, processing register template, AI impact assessment template
  • Service level agreement
  • Arabic-language support
  • Everything in Business

*Platform subscription gives access to Enclave42 features and model tiers. Inference usage is billed separately through your prepaid wallet.

PLATFORM ACCESS

Enterprise

Custom

For organisations that need the platform on their own or their partner's hardware.

Includes

  • Private or dedicated deployment, bring your own cloud
  • Platform licence on your infrastructure
  • Custom model access and fine-tunes
  • Custom SLA and commercial terms
  • Named technical contact
  • Custom analytics and reporting

*Platform subscription gives access to Enclave42 features and model tiers. Inference usage is billed separately through your prepaid wallet.