New Flagship Model Added: Qwen3.7-Max

Date: 2026-05-22 / (1405-03-01)

Summary

Alibaba's new flagship agent-foundation model Qwen3.7-Max is now available on AvalAI through the Chat Completions API (v1/chat/completions) with partial support via the Responses API (v1/responses). The model targets long-horizon agent workflows, frontier coding, and reasoning, and is launched with a 50% promotional discount through Jun 22.

---

Details

Alibaba

We announce access to Qwen3.7-Max (qwen3.7-max), Alibaba's latest proprietary model built for the agent era. The model excels at coding agents, office and productivity workflows, and sustained autonomous execution across hundreds or thousands of steps. Documentation

Qwen3.7-Max

Qwen3.7-Max is designed as a versatile agent foundation — equally capable of writing and debugging code, automating office workflows, and sustaining long-horizon autonomous execution. It generalizes across agent scaffolds such as Claude Code, OpenClaw, and Qwen Code, and supports hybrid thinking via DashScope's enable_thinking and preserve_thinking parameters.

Key Features:

• Endpoint Support: Available on v1/chat/completions (full feature compatibility) and partial support on v1/responses
• Agent Foundation: Frontier coding agent, office productivity, and long-horizon autonomous execution
• Cross-Harness Generalization: Consistent results across Claude Code, OpenClaw, Qwen Code, and custom tool-use frameworks
• Hybrid Thinking: Optional reasoning mode via extra_body={"enable_thinking": True} for streaming
• Preserve Thinking: Recommended for multi-turn agent workflows via preserve_thinking
• Strong Reasoning: 92.4 on GPQA Diamond, 97.1 on HMMT 2026 Feb, 90.0 on IMOAnswerBench
• Coding Excellence: 80.4 on SWE-Verified, 60.6 on SWE-Pro, 78.3 on SWE-Multilingual, 69.7 on Terminal-Bench 2.0
• Agent Benchmarks: 60.8 on MCP-Mark, 76.4 on MCP-Atlas, 87.0 on SpreadSheetBench-v1, 75.0 on BFCL-V4
• Long-Horizon Capability: Demonstrated 35-hour autonomous kernel optimization run with 1,158 tool calls and 10.0x speedup

Pricing Details (50% discount through Jun 22, 2026):

Model	Input	Cache Creation	Cached Input	Output
qwen3.7-max (promo)	$2.50 / 1M tokens	$3.125 / 1M tokens	$0.25 / 1M tokens	$7.50 / 1M tokens
qwen3.7-max (standard, after Jun 22)	$5.00 / 1M tokens	$6.25 / 1M tokens	$0.50 / 1M tokens	$15.00 / 1M tokens

---

API Request/Response Examples

Example Request

curl https://api.avalai.ir/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -d '{
    "model": "qwen3.7-max",
    "messages": [
      {
        "role": "user",
        "content": "Write a Python function to merge two sorted linked lists."
      }
    ]
  }'

Example Response

{
  "id": "chatcmpl-qwen37max-example",
  "created": 1779768420,
  "model": "qwen3.7-max",
  "object": "chat.completion",
  "system_fingerprint": null,
  "choices": [
    {
      "finish_reason": "stop",
      "index": 0,
      "message": {
        "content": "Here is a Python function that merges two sorted linked lists into a single sorted list using a dummy head pointer for clarity...\n\n
```python\nclass ListNode:\n    def __init__(self, val=0, next=None):\n        self.val = val\n        self.next = next\n\ndef merge_two_sorted_lists(l1, l2):\n    dummy = ListNode()\n    tail = dummy\n    while l1 and l2:\n        if l1.val <= l2.val:\n            tail.next, l1 = l1, l1.next\n        else:\n            tail.next, l2 = l2, l2.next\n        tail = tail.next\n    tail.next = l1 or l2\n    return dummy.next\n```",
        "role": "assistant",
        "thinking_blocks": [],
        "annotations": []
      }
    }
  ],
  "usage": {
    "completion_tokens": 175,
    "prompt_tokens": 22,
    "total_tokens": 197,
    "completion_tokens_details": null,
    "prompt_tokens_details": {
      "audio_tokens": null,
      "cached_tokens": null,
      "text_tokens": 22,
      "image_tokens": null
    }
  },
  "estimated_cost": {
    "unit": "0.0013675000",
    "irt": 156.72,
    "exchange_rate": 114600
  }
}

---

SDK Usage Examples

Bash

curl https://api.avalai.ir/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -d '{
    "model": "qwen3.7-max",
    "messages": [
      {
        "role": "user",
        "content": "Design a fault-tolerant message broker for a high-throughput trading system."
      }
    ]
  }'

Python

from openai import OpenAI

client = OpenAI(api_key="your-avalai-api-key", base_url="https://api.avalai.ir/v1")

completion = client.chat.completions.create(
    model="qwen3.7-max",
    messages=[
        {
            "role": "user",
            "content": "Design a fault-tolerant message broker for a high-throughput trading system.",
        }
    ],
)

print(completion.choices[0].message.content)

JavaScript

import { OpenAI } from "openai";

const client = new OpenAI({
  apiKey: process.env.AVALAI_API_KEY,
  baseURL: "https://api.avalai.ir/v1",
});

const completion = await client.chat.completions.create({
  model: "qwen3.7-max",
  messages: [
    {
      role: "user",
      content: "Design a fault-tolerant message broker for a high-throughput trading system.",
    },
  ],
});

console.log(completion.choices[0].message.content);

---

Hybrid Thinking with Streaming

Qwen3.7-Max supports DashScope's hybrid thinking mode. Use extra_body={"enable_thinking": True} together with stream=True to observe the reasoning content alongside the final answer. For multi-turn agent workflows, also enable preserve_thinking to keep prior reasoning available across turns:

Bash

curl https://api.avalai.ir/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -d '{
    "model": "qwen3.7-max",
    "messages": [
      {
        "role": "user",
        "content": "Plan a multi-step refactor that moves a monolithic payments service to event-driven microservices."
      }
    ],
    "stream": true,
    "enable_thinking": true,
    "preserve_thinking": true
  }'

Python

from openai import OpenAI

client = OpenAI(api_key="your-avalai-api-key", base_url="https://api.avalai.ir/v1")

stream = client.chat.completions.create(
    model="qwen3.7-max",
    messages=[
        {
            "role": "user",
            "content": "Plan a multi-step refactor that moves a monolithic payments service to event-driven microservices.",
        }
    ],
    stream=True,
    extra_body={
        "enable_thinking": True,
        "preserve_thinking": True,
    },
)

for chunk in stream:
    if not chunk.choices:
        continue
    delta = chunk.choices[0].delta
    if hasattr(delta, "reasoning_content") and delta.reasoning_content:
        print(delta.reasoning_content, end="", flush=True)
    if getattr(delta, "content", None):
        print(delta.content, end="", flush=True)

JavaScript

import { OpenAI } from "openai";

const client = new OpenAI({
  apiKey: process.env.AVALAI_API_KEY,
  baseURL: "https://api.avalai.ir/v1",
});

const stream = await client.chat.completions.create({
  model: "qwen3.7-max",
  messages: [
    {
      role: "user",
      content: "Plan a multi-step refactor that moves a monolithic payments service to event-driven microservices.",
    },
  ],
  stream: true,
  // @ts-ignore - DashScope-specific parameters passed through
  enable_thinking: true,
  preserve_thinking: true,
});

for await (const chunk of stream) {
  const delta = chunk.choices?.[0]?.delta;
  if (delta?.reasoning_content) process.stdout.write(delta.reasoning_content);
  if (delta?.content) process.stdout.write(delta.content);
}

Note

For non-streaming requests, set enable_thinking: false in extra_body. Hybrid thinking is only supported with streaming, per Alibaba's DashScope guidelines. See Alibaba Models for details.

---

Function Calling

Qwen3.7-Max supports function calling for agent workflows that need to connect model reasoning with external tools and systems:

Bash

curl https://api.avalai.ir/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -d '{
    "model": "qwen3.7-max",
    "messages": [
      {
        "role": "user",
        "content": "Check the production deploy status for service `payments-api` and report any failed pods."
      }
    ],
    "tools": [
      {
        "type": "function",
        "function": {
          "name": "get_deploy_status",
          "description": "Get current deployment status for a service",
          "parameters": {
            "type": "object",
            "properties": {
              "service": {
                "type": "string",
                "description": "Service name"
              }
            },
            "required": ["service"]
          }
        }
      }
    ],
    "tool_choice": "auto"
  }'

Python

tools = [
    {
        "type": "function",
        "function": {
            "name": "get_deploy_status",
            "description": "Get current deployment status for a service",
            "parameters": {
                "type": "object",
                "properties": {
                    "service": {
                        "type": "string",
                        "description": "Service name",
                    }
                },
                "required": ["service"],
            },
        },
    }
]

response = client.chat.completions.create(
    model="qwen3.7-max",
    messages=[
        {
            "role": "user",
            "content": "Check the production deploy status for service `payments-api` and report any failed pods.",
        }
    ],
    tools=tools,
    tool_choice="auto",
)

JavaScript

const tools = [
  {
    type: "function",
    function: {
      name: "get_deploy_status",
      description: "Get current deployment status for a service",
      parameters: {
        type: "object",
        properties: {
          service: {
            type: "string",
            description: "Service name",
          },
        },
        required: ["service"],
      },
    },
  },
];

const response = await client.chat.completions.create({
  model: "qwen3.7-max",
  messages: [
    {
      role: "user",
      content: "Check the production deploy status for service `payments-api` and report any failed pods.",
    },
  ],
  tools,
  tool_choice: "auto",
});

---

Responses API (Partial Support)

Qwen3.7-Max can also be invoked through the OpenAI-compatible Responses API (v1/responses). Partial support is available for text input/output and basic tool use; advanced built-in tools (e.g., web search, file search) and the reasoning configuration field remain limited to OpenAI's o-series models:

curl https://api.avalai.ir/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -d '{
    "model": "qwen3.7-max",
    "input": "Summarize the architectural risks of moving a SQL-based ledger to an event-sourced design."
  }'

See the Responses API reference for the full request and response schema.

---

Related Links

• Alibaba Models Documentation
• Reasoning Models Guide
• Chat Completions API
• Responses API
• Pricing Details
• Model Index