235B
Parameters
MoE (22B active)
AIVory Smart Inference · Qwen
Open WeightQwen 3 235B - frontier reasoning, open weights.
Alibaba's largest open model. 235B total parameters, 22B active per token. Competes with proprietary frontier models on reasoning benchmarks.
At a glance
Qwen 3 235B via Smart Inference.
Routed to the cheapest live provider on every request. Same model, lower bill.
128K
Context window
tokens per request
—
Cheapest input price
per 1M tokens via Smart Inference
—%
Cheaper
vs the priciest endpoint for the same model
Pricing
Qwen 3 235B: price comparison.
Our cheapest routed price vs what you'd pay elsewhere.
Prices in USD per 1M input tokens · updated live from the router
Quick start
Use Qwen 3 235B in three lines.
OpenAI-compatible. Swap your base_url, keep everything else.
from openai import OpenAI
client = OpenAI(
base_url="https://smart.aivory.net/v1",
api_key="sk-aivory-...",
)
resp = client.chat.completions.create(
model="qwen-3-235b",
messages=[{"role": "user", "content": "Ship it."}],
)
print(resp.choices[0].message.content)import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://smart.aivory.net/v1",
apiKey: process.env.AIVORY_API_KEY,
});
const resp = await client.chat.completions.create({
model: "qwen-3-235b",
messages: [{ role: "user", content: "Ship it." }],
});
console.log(resp.choices[0].message.content);curl https://smart.aivory.net/v1/chat/completions \
-H "Authorization: Bearer $AIVORY_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "qwen-3-235b",
"messages": [{"role": "user", "content": "Ship it."}]
}'That's the whole migration. Get early access and try it →
Use cases
What Qwen 3 235B is great at.
-
Frontier-Quality Reasoning
Qwen 3 235B scores within range of GPT-4o and Claude 3.5 Sonnet on MMLU, MATH, and coding benchmarks. Use it when you need open-weight and the best available reasoning quality in the same package.
-
Agentic Tool-Use Chains
Qwen 3 is tuned for multi-step tool calling. Give it a set of functions, a goal, and let it plan and execute. The 22B active parameters keep per-step cost low while the full 235B parameter bank gives it enough depth for planning.
-
Long Research Synthesis
Feed 100K tokens of source material and ask for a structured analysis. Qwen 3 maintains coherence across long outputs better than most open models. The 128K context window fits substantial research corpora.
FAQ
Common questions.
What does Qwen 3 235B cost on AIVory?
More than a 70B model, less than you'd expect for a 235B one. The MoE architecture (22B active) keeps inference cost moderate. Check the live table above for current per-token rates.
Qwen 3 235B vs DeepSeek V3.2 - how do they compare?
Both are frontier MoE models. V3.2 has more total parameters (671B vs 235B) but activates more per token (37B vs 22B), making it slightly more expensive. Quality is close on most benchmarks. Qwen 3 tends to be faster due to the smaller active count.
Is Qwen 3 good for coding?
Yes, competitive on HumanEval and MBPP. But if coding is your primary use case, consider Qwen 3.5 27B first. It's specifically tuned for agentic coding and costs a fraction of the 235B model.
What are the commercial terms?
Alibaba released Qwen 3 under the Qwen License. Commercial use is allowed with no royalties. The license is more permissive than Meta's Llama license.
Start using Qwen 3 235B. Today.
Swap your base_url, keep shipping.
The router is live. Credits from $10, no subscription, no commitment.