Private Models

Qwen: Qwen3.5 397B A17B

Trusted Execution Environment
qwen/qwen3.5-397b-a17bModel weights

The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. It delivers state-of-the-art performance comparable to leading-edge models across a wide range of tasks, including language understanding, logical reasoning, code generation, agent-based tasks, image understanding, video understanding, and graphical user interface (GUI) interactions. With its robust code-generation and agent capabilities, the model exhibits strong generalization across diverse agent.

Modalities
TextImageVideotoText
In / out price
$0.55 / $3.50per 1MCache read $0.225 per 1M
Context
262K
Released
Feb 28, 2026

Providers

RedPill routes this private model across verified providers with automatic fallbacks. Pricing is unified, so you pay the same price no matter which provider serves your request, while live metrics show first-token latency, speed, and uptime.

chutesCheckingCheckingChecking

Start building

Call this model from your application, or through your agents.

For applications

Point your OpenAI SDK at the Gateway. Chat, completions, and embeddings work unchanged.

1
Create a key
2
Set base URL
3
Verify receipt

For agents

Connect Claude Code, Codex, Pi, and more through Private AI Proxy. Traffic is protected while a verified session is active.

1
Install Proxy
2
Turn on Protected
3
Connect agents
Release notes
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://tee.redpill.ai/v1",
  apiKey: process.env.REDPILL_API_KEY,
});

const response = await client.chat.completions.create({
  model: "qwen/qwen3.5-397b-a17b",
  messages: [{ role: "user", content: "Say hello." }],
});

Check privacy evidence

Follow the receipt from the gateway response to its verified session using four curl requests.

  1. 1
    Check the gateway

    Fetch the current TEE evidence and check that the keyset has not expired.

    curl -fsS "https://tee.redpill.ai/v1/aci/attestation" \
      -H "Authorization: Bearer $REDPILL_API_KEY"

    Output

    {
      "api_version": "aci/1",
      "workload_keyset_digest": "sha256:4f3dc45e...",
      "attestation": { "workload_keyset": { "not_after": 1790498751, ... }, ... },
      ...
    }
  2. 2
    Send a private request

    The response header contains the receipt ID, not the receipt itself.

    curl -fsS -i "https://tee.redpill.ai/v1/chat/completions" \
      -H "Authorization: Bearer $REDPILL_API_KEY" --json '{"model":"qwen/qwen3.5-397b-a17b","messages":[{"role":"user","content":"Hello"}],"provider":{"aci_verified":true}}'

    Output

    HTTP/2 200
    x-receipt-id: rcpt-e0eefe63b3673e2a5984951f
    ...
    
    {
      "model": "qwen/qwen3.5-397b-a17b",
      ...
    }
  3. 3
    Inspect the receipt

    Paste the x-receipt-id above and compare its keyset digest with step 1.

    curl -fsS "https://tee.redpill.ai/v1/aci/receipts/REPLACE_WITH_RECEIPT_ID" \
      -H "Authorization: Bearer $REDPILL_API_KEY"

    Output

    {
      "api_version": "aci/1",
      "receipt_id": "rcpt-e0eefe63b3673e2a5984951f",
      "workload_keyset_digest": "sha256:4f3dc45e...",
      "event_log": [
        ...,
        { "type": "upstream.verified", "session_id": "63c5a08de14e5c5253e6fb29271151f40621beb83b552ccca65772772856053f", "result": "verified", "required": true },
        { "type": "response.returned", "body_hash": "sha256:..." }
      ],
      ...
    }
  4. 4
    Inspect the session

    Copy session_id from the upstream.verified event above, then fetch that session.

    curl -fsS "https://tee.redpill.ai/v1/aci/sessions/PASTE_SESSION_ID_FROM_RECEIPT" \
      -H "Authorization: Bearer $REDPILL_API_KEY"

    Output

    {
      "api_version": "aci/1",
      "upstream_name": "...",
      "verifier_id": "...",
      "claims": { "tee_attested": { "status": "asserted", ... }, ... },
      ...
    }
Full cryptographic verification guide

Search

Search models and pages.