Alibaba

Qwen3-VL-235B-A22B-Thinking

Text

Retired Retired on 2026-10-10.

Thinking-only Qwen3-VL vision MoE; 131K context; image input via OpenAI-compatible API

Retired Qwen3-VL-235B-A22B-Thinking API reference →

Get API Key

Model health Insufficient data

runapi.ai

# Base URL
https://runapi.ai

# Endpoints
POST /v1/chat/completions
POST /v1/responses
POST /v1/messages
POST /v1beta/models/qwen3-vl-235b-a22b-thinking:generateContent
POST /v1beta/models/qwen3-vl-235b-a22b-thinking:streamGenerateContent

curl https://runapi.ai/v1/chat/completions \
  -H "Authorization: Bearer $RUNAPI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "qwen3-vl-235b-a22b-thinking",
  "messages": [
    {
      "role": "user",
      "content": "Read this screenshot of a failing dashboard, explain what the error message means, and suggest the code change that would fix it."
    }
  ]
}'

from openai import OpenAI

client = OpenAI(
    base_url="https://runapi.ai/v1",
    api_key="your-runapi-key"
)

response = client.chat.completions.create(
    model="qwen3-vl-235b-a22b-thinking",
    messages=[{"role": "user", "content": "Read this screenshot of a failing dashboard, explain what the error message means, and suggest the code change that would fix it."}]
)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://runapi.ai/v1",
  apiKey: "your-runapi-key"
});

const response = await client.chat.completions.create({
  model: "qwen3-vl-235b-a22b-thinking",
  messages: [{ role: "user", content: "Read this screenshot of a failing dashboard, explain what the error message means, and suggest the code change that would fix it." }]
});

https://runapi.ai 5 endpoints

9 versions available

OVERVIEW

What Qwen3-VL-235B-A22B-Thinking is built for

Qwen3-VL-235B-A22B-Thinking is the reasoning version of Qwen3-VL, the flagship vision-language model family from Alibaba Qwen. It is a Mixture-of-Experts model with 235B total parameters and 22B active per token, released with open weights under Apache 2.0 on September 23, 2025. Alibaba tunes it for multimodal reasoning, with a focus on STEM and math problems, OCR, and image and video understanding.

The model is thinking-only: every response starts with a reasoning step, and the step cannot be switched off. Alibaba's API serves it with a 131,072-token context window and up to 32,768 output tokens.

On RunAPI, call qwen3-vl-235b-a22b-thinking through the OpenAI-compatible Chat Completions endpoint (/v1/chat/completions) with an OpenAI SDK pointed at https://runapi.ai/v1. Send images as image_url parts next to the text in messages[].content[]. Alibaba now lists Qwen3-VL among its previous-generation models, because the Qwen3.5 and Qwen3.6 models read images natively.

Model ID

qwen3-vl-235b-a22b-thinking

Released

September 23, 2025

Architecture

MoE, 235B total / 22B active, Apache 2.0 weights

Context window

131,072 tokens

Max output

32,768 tokens

Thinking

Always on

Endpoint on RunAPI

OpenAI Chat Completions, text and image input

Specifications

Provider	Alibaba
Model ID	qwen3-vl-235b-a22b-thinking
Modality	Text
Task types	Synchronous
API endpoint	/v1/chat/completions
	/v1/responses

	/v1/messages
	/v1beta/models/qwen3-vl-235b-a22b-thinking:generateContent
	/v1beta/models/qwen3-vl-235b-a22b-thinking:streamGenerateContent
Billing unit	1K tokens
Catalog status	Retired

HOW IT WORKS

Use This Model in 3 Steps

Get Your Key

Sign up for free and create an API key for qwen3-vl-235b-a22b-thinking from the dashboard.
Send a Request

POST to /v1/chat/completions with the qwen3-vl-235b-a22b-thinking model slug and your parameters.
Get Results

Read the completed qwen3-vl-235b-a22b-thinking response directly from the endpoint.

What this model offers

Endpoint coverage

Exposes 1 public endpoint(s): chat_completion.

Usage-based billing

Metered by 1K tokens with no subscription requirement.

API REFERENCE

Endpoint Specification

Endpoint: POST /v1/chat/completions
Model ID: qwen3-vl-235b-a22b-thinking
Task Types: chat_completion

Request parameters

No request parameters are published for this endpoint.

USE CASES

Customer support

Answer customer questions from a private knowledge base, reducing ticket volume.

Document analysis

Draft contract summaries and flag key clauses for attorney review.

Code generation

Auto-generate unit tests, code reviews, and refactoring suggestions in CI.

FAQ

Frequently asked questions about Qwen3-VL-235B-A22B-Thinking

What is the difference between the Thinking and Instruct versions of Qwen3-VL-235B-A22B?

Both use the same 235B Mixture-of-Experts model with 22B active parameters. The Thinking version always reasons before it answers and is tuned for STEM, math, and multimodal reasoning. The Instruct version answers directly and is aimed at visual coding, spatial perception, long-video understanding, and OCR.

How do I send an image to qwen3-vl-235b-a22b-thinking?

Use the Chat Completions endpoint and add an image_url part next to the text part in messages[].content[], the same request shape the OpenAI SDK uses for vision. Alibaba also lists video input for this model; video requests are not verified on RunAPI.

Can I turn thinking off on this model?

No. Alibaba ships it as a thinking-only model. If you want answers without a reasoning step, call qwen3-vl-235b-a22b-instruct instead.

Should I use Qwen3-VL or a Qwen3.5 model for image tasks?

Alibaba lists Qwen3-VL among its previous-generation models and says the vision capability of Qwen3.5-122B-A10B exceeds Qwen3-VL-235B. For new image workloads, try a Qwen3.5 or Qwen3.6 model first, and keep Qwen3-VL for prompts and evaluations already built on it.

How do I select qwen3-vl-235b-a22b-thinking?

Pass the model ID shown in the quickstart.

What's the rate limit on this variant?

Per-key rate limits scale with your usage tier. The pricing page shows current limits. If you need higher throughput, contact support to discuss tier upgrades.

Can I switch variants later?

Yes. Variant is a parameter in the request. Switch by changing the model ID — no code changes, no re-authentication, no separate billing setup. All variants share the same API key and request shape.

Does it stream?

Where streaming is available, RunAPI streams end-to-end. LLM models support token-level streaming. Media models use async task polling or webhook callbacks for result delivery.

Where do I report quality issues?

Open an issue on the public GitHub repo or email support at [email protected]. Include the task ID and model ID so the team can investigate the specific generation.

Do I need a separate provider account?

No. Your RunAPI API key is enough to access this variant and every other model in the catalog. You do not need accounts with the underlying provider.

Building with a team?

We're here to help with enterprise setup, integrations, and technical questions.