Alibaba (Qwen)

Qwen3 VL 235B A22B Thinking

Generally availableOpen weights131K contextReasoning

Summary

Qwen3 VL 235B A22B Thinking is an open-weight language model from Alibaba (Qwen), released on 23 Sept 2025. It accepts text and images and generates text, with a 131K-token context window and up to 33K output tokens per response. As of 18 Nov 2026, the lowest listed API price is $0.98 per million input tokens and $3.95 per million output tokens (Hugging Face Inference Providers), across 1 provider we track. It is scheduled for retirement on 9 Oct 2026.

Qwen vision-language model for visual reasoning, documents, and agent tasks

Last verified 18 Nov 2026. Every figure on this page links to its source.

Specifications

DeveloperAlibaba (Qwen)
Release date23 Sept 2025 [models.dev]
StatusGenerally available [models.dev]
WeightsOpen weights [Hugging Face]
Licenseapache-2.0 [Hugging Face]
Parameters236B [Hugging Face]
Context window131K (131,072 tokens) [models.dev]
Max output33K tokens [models.dev]
Knowledge cutoff31 Mar 2025 [models.dev]
InputText and images [models.dev]
OutputText [models.dev]
Reasoning modeYes [models.dev]
Tool callingYes [models.dev]
Structured outputYes [models.dev]
Retirement date9 Oct 2026 [OpenRouter]
Hugging FaceQwen/Qwen3-VL-235B-A22B-Thinking [Hugging Face]

Qwen3 VL 235B A22B Thinking API pricing by provider

USD per million tokens. Sorted by input price.

ProviderInputOutputCached inputContextSource
Hugging Face Inference Providers
Qwen/Qwen3-VL-235B-A22B-Thinking
$0.98$3.95–131Kmodels.dev, 18 Nov 2026

Running Qwen3 VL 235B A22B Thinking locally

Estimated memory, from the parameter count and the architecture in config.json on Hugging Face. Includes about 0.8 GB of runtime overhead.

QuantizationWeightsTotal, 8K contextTotal, 32K context
Q4_K_M (4-bit)143 GB145 GB150 GB
Q8_0 (8-bit)250 GB253 GB258 GB
FP16 / BF16 (16-bit)471 GB474 GB478 GB

At 4-bit with an 8K conversation, Qwen3 VL 235B A22B Thinking needs about 145 GB, more than any single consumer graphics card; it needs a unified-memory machine, a multi-GPU setup or a split between graphics card and system RAM. Check your own hardware, or see Best computer for local AI on our sister site AIPCs.

Frequently asked questions

What is the context window of Qwen3 VL 235B A22B Thinking?

Qwen3 VL 235B A22B Thinking has a context window of 131,072 tokens (131K) and can generate up to 32,768 tokens in a single response, according to models.dev.

How much does the Qwen3 VL 235B A22B Thinking API cost?

Qwen3 VL 235B A22B Thinking is listed from $0.98 per million input tokens and $3.95 per million output tokens (Hugging Face Inference Providers). Prices last verified 18 Nov 2026.

Is Qwen3 VL 235B A22B Thinking open source?

Qwen3 VL 235B A22B Thinking is an open-weight model: its weights are published on Hugging Face as Qwen/Qwen3-VL-235B-A22B-Thinking under the apache-2.0 license. Check the license terms before commercial use.

When was Qwen3 VL 235B A22B Thinking released?

Alibaba (Qwen) released Qwen3 VL 235B A22B Thinking on 23 Sept 2025, according to models.dev.

How much VRAM does Qwen3 VL 235B A22B Thinking need?

About 145 GB at 4-bit (Q4_K_M) with an 8K-token conversation, and about 253 GB at 8-bit. That is more than any single consumer graphics card holds. These are estimates from the model's published parameter count and architecture.

Similar models from other labs

Comparable on output type, weights, release date, price and context window. Prices are the lowest listed, USD per million tokens.

ModelLabReleasedContextInput $/1MOutput $/1MWeights
Hermes-4-405BNousresearch26 Aug 2025131K$1$3Open
GLM-4.5Zhipu AI (Z.ai)28 Jul 2025131K$0.60$2.20Open
Magistral Small 1.2Mistral AI18 Sept 2025128K$0.50$1.50Open
Kimi K2 InstructMoonshot AI11 Jul 2025131K$0.57$2.30Open
Kimi-K2-InstructMoonshot AI11 Jul 2025131K$0.57$2.30Open
GLM-4.6Zhipu AI (Z.ai)30 Sept 2025205K$0.50$2Open

Change history

  • No longer listed on openrouter
  • No longer listed on novita-ai

Sources