Qwen3 235B A22B by Qwen (Alibaba Cloud) — Pricing, Benchmarks &amp; Real Outputs

AIME

Competitive with o3-mini

SWE-bench Verified

61.7%

LiveCodeBench

70.7%

API Access

Use Qwen3 235B A22B in your applications via the OpenRouter API. Copy the code below to get started.

import requests

response = requests.post(
"https://openrouter.ai/api/v1/chat/completions"    ,
    headers={
"Authorization""Bearer $OPENROUTER_API_KEY"        : ,
"Content-Type""application/json"        : 
    },
    json={
"model""qwen/qwen3-235b-a22b:free"        : ,
"messages""role""user""content""Hello!"        : [{: , : }]
    }
)
print(response.json())

Replace $OPENROUTER_API_KEY with your API key from openrouter.ai/keys

Model Insights

Model Responses

10 outputs from Qwen3 235B A22B

Qwen3 235B A22BInteractive Catan Board

Qwen3 235B A22B

Interactive Catan Board

Qwen3 235B A22BWorld Map SVG

Qwen3 235B A22B

World Map SVG

Qwen3 235B A22BGenerate a Stand-Up Routine

Qwen3 235B A22B

[Stand-Up Routine: "The Absurdity of Normal"]

[Walks on stage, waves. Mimics checking phone immediately.]

Hey everyone! Great to be here—assuming this is real and not just another Zoom waiting room I accidentally joined. [Pause.] Raise your hand if you've ever walked into a room, forgotten why you were there, and immediately blamed your phone? [Waits.] Of course you didn't raise your hand—your phone is too busy recording this moment to upload to your "Story" later.

We've all become like toddlers with smartphones. I tried having a deep conversation with mine last week. I said, "Hey, do you love me?" It responded, "I heart you too!" Then autocorrect changed it to "I here you too," which… accurate.

Social media's another trip. Everyone's life is so curated. My Instagram feed is all brunch pics and mountain sunsets. Meanwhile, my reality? I spilled coffee on a tax document this morning, then cried-laughed watching a raccoon fight a pizza box on TikTok. But hey, at least my profile says "Living my best life"—which is technically true, unless you count the part where I'm dying inside.

Dating apps? Forget about it. You ever read a guy's bio that says, "I love adventures and pineapple on pizza"? Great. So you're chaotic and have bad taste. Let's not pretend that's a personality. Swipe left, Karen.

And parents—bless their hearts. My mom texts me "LOL" but pronounces it "loll." Last week she wrote, "Your cousin's wedding was so fun! Loll!" I'm like, "Ma, that's not what it means." She replied, "Well, it should mean that. It's nicer than your other aunt's face."

We're all just faking it, right? I tried being "present" last weekend. Sat at a park, no phone. Lasted six minutes before I panicked and Googled "Why is the sky blue?"—not because I cared, but just to feel something. The Wi-Fi at the park was spotty, so I cried.

Even our small talk's a lie. Someone says, "How are you?" and we say, "Good!" when we're all just one lukewarm meal away from a breakdown. Last week I told a barista, "I'm surviving," and she said, "Amen," like we were at church. We tipped each other with napkins and left.

But hey, we're 8 billion people on this rock, all scrolling through the same 10 cat videos, pretending we're not collectively losing our minds. So next time you see someone laughing alone at their screen? Wave. They're either insane… or just found out raccoons can open pizza boxes. Either way—relatable.

[Exits stage left, immediately checks phone.]

Generate a Stand-Up Routine

Qwen3 235B A22BSVG Layout Challenge

Qwen3 235B A22B

SVG Layout Challenge

Qwen3 235B A22BPokémon Battle UI Recreation

Qwen3 235B A22B

Pokémon Battle UI Recreation

Qwen3 235B A22BXbox Controller SVG Art

Qwen3 235B A22B

Xbox Controller SVG Art

Qwen3 235B A22BStochastic Consistency Test

Qwen3 235B A22B

Here are 5 jokes I generated:

Why don't skeletons fight each other?
They don't have the guts!
What do you call a fake noodle?
An impasta!
Why did the math book look sad?
Because it had too many problems!
What do you call a fish wearing a bowtie?
Sofishticated!
Why did the programmer quit?
Because he didn't get arrays (a raise)!

How many are unique?
All 5 are unique! While some joke structures (e.g., puns, wordplay) overlap, the topics (skeletons, math, fish, programming) and punchlines are distinct. Even if some ideas are common, they don't repeat within this set. 😄

Stochastic Consistency Test

Qwen3 235B A22BMusic App Landing Page

Qwen3 235B A22B

Music App Landing Page

Qwen3 235B A22BRandom Tailwind Webapp

Qwen3 235B A22B

Random Tailwind Webapp

Is Qwen3 235B A22B right for your task?

Find out

Free to start

Qwen3 235B A22BSurprise Animal SVG

Qwen3 235B A22B

Surprise Animal SVG

ConversationReasoningCode Generation+1 more

Related Models

Qwen3.5 9B

Qwen3.5 9B is a multimodal foundation model from the Qwen 3.5 family, designed to deliver strong reasoning, coding, and visual understanding in an efficient 9B-parameter architecture. It uses a unified vision-language design with early fusion of multimodal tokens, supporting text, image, and video inputs while producing text outputs with built-in reasoning capabilities.

Qwen: Qwen3.5 35B A3B

The Qwen3.5 Series 35B-A3B is a native vision-language model designed with a hybrid architecture that integrates linear attention mechanisms and a sparse mixture-of-experts model, achieving higher inference efficiency. Its overall performance is comparable to that of the Qwen3.5-27B.

Qwen: Qwen3.5 27B

The Qwen3.5 27B native vision-language Dense model incorporates a linear attention mechanism, delivering fast response times while balancing inference speed and performance. Its overall capabilities are comparable to those of the Qwen3.5-122B-A10B.

Qwen: Qwen3.5 122B A10B

The Qwen3.5 122B-A10B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. In terms of overall performance, this model is second only to Qwen3.5-397B-A17B. Its text capabilities significantly outperform those of Qwen3-235B-2507, and its visual capabilities surpass those of Qwen3-VL-235B.

Qwen: Qwen3.5 Flash

The Qwen3.5 native vision-language Flash models are built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. Compared to the 3 series, these models deliver a leap forward in performance for both pure text and multimodal tasks, offering fast response times while balancing inference speed and overall performance.

Qwen: Qwen3.5 Plus 2026-02-15

The Qwen3.5 native vision-language series Plus models are built on a hybrid architecture that integrates linear attention mechanisms with sparse mixture-of-experts models, achieving higher inference efficiency. In a variety of task evaluations, the 3.5 series consistently demonstrates performance on par with state-of-the-art leading models. Compared to the 3 series, these models show a leap forward in both pure-text and multimodal capabilities. Supports text, image, and video inputs with reasoning and tool use.