Qwen3 VL 235B A22B Instruct

qwen/qwen3-vl-235b-a22b-instruct
by Qwen · 2025-09-23

Qwen3-VL 235B-A22B Instruct — open-weight vision-language model, 235B total / 22B active params, 256k context, no thinking mode.

p50 TTFT10.00 s
INPUT$0.40/ 1M tokens
OUTPUT$1.60/ 1M tokens
p50 TTFT10.00 s7d
p95 TTFT10.00 s7d
TRAFFIC77.1Ktokens / 7d

Qwen3 VL 235B A22B Instruct is a vision-language variant of the Qwen3 model series, built by Alibaba Cloud. It uses a mixture-of-experts design with 235 billion total parameters, of which…

What exactly is Qwen3 VL 235B A22B Instruct?

Who is this model best suited for?

What modalities does the model support?

How does OrcaRouter deliver this model?

Code samples

Call from any SDK

OpenAI-compatible — keep the SDK you already use

  • OpenAI SDKhttps://api.orcarouter.ai/v1
from openai import OpenAI

client = OpenAI(
    base_url="https://api.orcarouter.ai/v1",
    api_key="$ORCAROUTER_API_KEY",
)

response = client.chat.completions.create(
    model="qwen/qwen3-vl-235b-a22b-instruct",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Pricing

Input / 1M tokens$0.400
Output / 1M tokens$1.60
CurrencyUSD

Cost calculator

Tokens / month10MM
Input share70%%
Estimated / month $7.60

Estimate based on list price

Token & cost estimator

Input tokens: 20Cost per request: $0.000808

Estimate only — actual token counts depend on the provider's tokenizer.

Performance

p50 TTFT
10.00 s
Output speed
70.3 tok/s
p95 TTFT
10.00 s
Error rate
0%

Public benchmarks

16.5
AA Coding
Better than 11% of models compared
#109 of 123
14.3
AA Intelligence
Better than 15% of models compared
#106 of 125
70.7
AA Math
Better than 64% of models compared
#29 of 81
AIME 2025
70.7
GPQA Diamond
71.2
Humanity's Last Exam
6.3
IFBench
42.7
LiveCodeBench
59.4
Long-Context Recall
31.7
MMLU-Pro
82.3
SciCode
35.9
TerminalBench Hard
6.8
τ²-Bench
35.1
Source: artificialanalysis.ai

How it compares

Qwen3 VL 235B A22B Instructqwen/qwen3-max-previewQwen3.5 397B A17Bqwen/qwen3.5-plus
Input $/M$0.40$0.86$0.17$0.12
Output $/M$1.60$3.44$1.03$0.69
Context262K33K1.0M
Quality8/108/108/10
Compare side-by-sideCompare side-by-sideCompare side-by-sideCompare side-by-side

FAQ

What is the cost in dollars per million tokens for this model on OrcaRouter?
Input tokens cost $0.40 per million and output tokens cost $1.60 per million. Image tokens are billed at the input rate. There is no additional markup.
What is the context window size of Qwen3 VL 235B A22B Instruct?
The exact context window length is not provided in the available facts. Typically, Qwen3 VL models support large context lengths (e.g., 32K tokens), but you should verify with the provider's official documentation.
What are the model's main strengths compared to other VL models?
Its MoE architecture provides a good balance of capacity and cost. The MMLU-Pro score of 82.3 indicates strong general knowledge. It is also one of the most affordable large multimodal models available through OrcaRouter.
How does this model compare to GPT-4V in terms of performance?
Direct benchmark comparisons are not provided, but Qwen3 VL's MMLU-Pro score is 82.3. GPT-4V's performance on the same benchmark is not publicly standardized. Pricing favors Qwen3 VL, which is significantly cheaper per token.
Does OrcaRouter handle my data when I send images or text?
OrcaRouter's data handling policies are not detailed here. Generally, the gateway passes data to the provider. You should review OrcaRouter's privacy policy and terms of service for specifics on data storage, retention, and sharing.
How can I call this model via an OpenAI-compatible API?
Use the base URL https://api.orcarouter.ai/v1 with model ID "qwen/qwen3-vl-235b-a22b-instruct". Send a POST request to /chat/completions with a messages array containing text and image content. Use your OrcaRouter API key for authentication.
Is there any free tier or trial for using this model on OrcaRouter?
The available facts do not mention a free tier. OrcaRouter likely charges per usage. You should check their website for current offers or trial credits.
What image formats are supported and how are they billed?
The model accepts common image formats such as JPEG and PNG (exact list from provider). Images are converted to tokens based on resolution and detail setting. Those tokens count toward input at $0.40 per million tokens.
Can I use this model for real-time applications?
Latency is higher than smaller models due to its 235B total parameters. It may not be suitable for very low-latency real-time use. For near-real-time multimodal inference, consider using a smaller model available on OrcaRouter.
Does this model support multi-image input?
Yes, Qwen3 VL models generally support multiple images in a single conversation. You can include several image URLs or base64 data in one message or across multiple turns.

Embed this badge

Qwen: Qwen3 VL 235B A22B Instruct$0.40/M in10000ms p50via OrcaRouter
HTML <a href="https://www.orcarouter.ai/models/qwen/qwen3-vl-235b-a22b-instruct" target="_blank"> <img src="https://www.orcarouter.ai/embed/qwen/qwen3-vl-235b-a22b-instruct.svg" alt="Qwen: Qwen3 VL 235B A22B Instruct on OrcaRouter" /> </a>
Markdown [![Qwen: Qwen3 VL 235B A22B Instruct](https://www.orcarouter.ai/embed/qwen/qwen3-vl-235b-a22b-instruct.svg)](https://www.orcarouter.ai/models/qwen/qwen3-vl-235b-a22b-instruct)

Model card as data

GET /api/public/models/qwen/qwen3-vl-235b-a22b-instructOpen
Machine-readable:/llms.txt/llms-full.txt