Hy3 · aihubmix

The Hy3 official version is honed for real-world business scenarios, using a Mixture-of-Experts (MoE) architecture with 295B total parameters and 21B activated parameters. It natively supports a 256K context window and offers multiple thinking modes: no_think (ultra-fast response), think_low (quick thinking), and think_high (deep reasoning), balancing ultra-fast responses, complex reasoning, and invocation cost. Compared with the Preview version, Hy3—based on real business feedback from Tencent Yuanbao, WorkBuddy, ima, Marvis, and others—focuses on improving the Coding Agent, long-form understanding, multi-turn context continuity, search QA, and complex task execution, performing more stably in reducing hallucinations, improving task completion, and engineering usability. It is better suited to practical scenarios such as frontend tasks, cross-file code development, long-document analysis, office automation, and multi-step Agent workflows.

Basismodell Alle Modelle dieser Basis Alle Modelle des Anbieters zum Anbieter

Eingabe, $/1M $0.16

Ausgabe, $/1M $0.62

Cache-Lesen $0.03905

Cache-Schreiben —

Kontext 256 000

Max. Ausgabe 128 000

Akzeptiert Text

Zuletzt geprüft 04.11.2026 12:00:14 UTC

Quelle Seite des Anbieters

Preisalarme

Anmelden, um dem Preis zu folgen

API

OpenAI-kompatible API: Die offiziellen OpenAI-SDKs funktionieren mit dieser Basis-URL. Ersetzen Sie YOUR_API_KEY durch Ihren Schlüssel vom Anbieter.

API-Basis-URL	`https://aihubmix.com/api/v1`
Endpunkt	`https://aihubmix.com/api/v1/chat/completions`
Modell-ID	`hy3`

curl

curl https://aihubmix.com/api/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "hy3", "messages": [{"role": "user", "content": "Hello!"}], "max_tokens": 256}'

Python

import requests

r = requests.post(
    "https://aihubmix.com/api/v1/chat/completions",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    json={
      "model": "hy3",
      "messages": [
        {
          "role": "user",
          "content": "Hello!"
        }
      ],
      "max_tokens": 256
    },
)
print(r.json())

JavaScript

const r = await fetch("https://aihubmix.com/api/v1/chat/completions", {
  method: "POST",
  headers: {Authorization: "Bearer YOUR_API_KEY", "Content-Type": "application/json"},
  body: JSON.stringify({
    "model": "hy3",
    "messages": [
      {
        "role": "user",
        "content": "Hello!"
      }
    ],
    "max_tokens": 256
  }),
});
console.log(await r.json());

Anfrage (JSON)

{
  "model": "hy3",
  "messages": [
    {
      "role": "user",
      "content": "Hello!"
    }
  ],
  "max_tokens": 256
}

Antwort (JSON)

{
  "id": "chatcmpl-...",
  "object": "chat.completion",
  "model": "hy3",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hi! How can I help?"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 9,
    "completion_tokens": 12,
    "total_tokens": 21
  }
}

Die Adresse stammt aus der Modellliste des Anbieters; bitte mit seiner Dokumentation abgleichen.

API-Dokumentation des Anbieters

Source: AI Rate Card, https://airatecard.com/de/c/offer%3Aaihubmix%7Chy3, prices checked 2026-11-04; seller's page: https://aihubmix.com

Technische Details

ID: offer:aihubmix|hy3