impulse

Qwen 3 32B

Open weights

Qwen3-32B is a dense 32.8B parameter causal language model from the Qwen3 series, optimized for both complex reasoning and efficient dialogue.

API id qwen-3-32bReleased 28 Apr 2025 (Apr 2025)
Input / 1M
$0.10
at Vercel AI Gateway
Output / 1M
$0.30
at Vercel AI Gateway
Context window
41K
tokens
Max output
16K
tokens per reply
Hosts
3
3 listings
Released
Apr 2025
28 Apr 2025

Pricing

List prices at Vercel AI Gateway, in US dollars per million tokens.

TierInput / 1MOutput / 1M
Standard$0.10$0.30

What it costs a month

At Vercel AI Gateway list prices, 30 days.

  • Chat assistant5,000 requests a day, 2K in / 0.5K out$52.50
  • RAG over documents2,000 requests a day, 8K in / 0.6K out$58.80
  • Coding agent1,000 requests a day, 40K in / 2K out$138
  • Overnight batch20,000 requests a day, 4K in / 0.4K out$312
Use your own numbers

Where to run it

3 hosts serve Qwen 3 32B. The comparison is on blended price, 3 parts input to 1 part output.

HostIn / out per 1M
qwen-3-32b
$0.10 / $0.30
qwen-3-32b
$0.108 / $0.378
+17% vs headline
CerebrasInference
qwen-3-32b
$0.40 / $0.80
+233% vs headline

Capabilities

Core

  • Streaming
  • Function calling
  • JSON mode
  • Structured outputs

Multimodal

  • Vision
  • Image input
  • Audio
  • Image generation
  • Video

Advanced

  • Reasoning
  • Tool calling
  • MCP
  • Prompt caching
  • Batch API
  • Embeddings
Takes in
Text
Produces
Text
Open weights
License: Apache-2.0
Runs in about 78.6 GB of GPU memory
Runtimes: Ollama, SGLang, llama.cpp, vLLM

Cheaper alternatives

Current models that cost less and keep most of what this one can do: tools, vision, reasoning, structured output, caching and enough context.

Same model, cheaper host

No host sells it for less than Vercel AI Gateway.

Smaller Alibaba Qwen models

No cheaper Alibaba Qwen model keeps its key capabilities.

Other frontier labs

No cheaper match from another frontier lab.

Open-weight picks

  • Ministral 8B 2512
    Mistral AI · on Mistral AI
    $0.15 / $0.15
    Has every key capability
    0% less
  • DeepSeek V4.1 Flash
    DeepSeek · on Sail
    $0.09 / $0.30
    Has every key capability
    5% less
  • MiMo V2.6 Flash
    Xiaomi · on GMI Cloud
    $0.112 / $0.224
    Has every key capability
    7% less
  • Qwen3.6 35B A3B
    Alibaba Qwen · on Sail
    $0.05 / $0.40
    Has every key capability
    8% less

Benchmarks

Scores published on the model card.

  • Aider Polyglot40.0%

Model line

Newer and older versions, at each lab's own price.

  1. Apr 2025Qwen3 32BLatest$0.05 / $0.10
  2. Apr 2025Qwen 3 32BThis page$0.10 / $0.30
  3. UndatedQwen2.5 32B$0.90 / $0.90

Call it

Using Vercel AI Gateway's API.

Python
from litellm import completion

response = completion(
    model="vercel_ai_gateway/qwen-3-32b",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Quick answers

How much does Qwen 3 32B cost?

On Vercel AI Gateway, Qwen 3 32B costs $0.10 per million input tokens and $0.30 per million output tokens.

Which APIs offer Qwen 3 32B?

Qwen 3 32B runs on 3 APIs: Vercel AI Gateway, Kyma API and Cerebras.

What is Qwen 3 32B's context window?

41K tokens, with up to 16K tokens in one reply.

Is Qwen 3 32B being retired?

No retirement date is scheduled for it on any API we track.

Is there a newer version of Qwen 3 32B?

Yes. Qwen3 32B is the newest model in this line, released on 28 Apr 2025.

What are cheaper alternatives to Qwen 3 32B?

Models that cost less and keep most of what it can do: Ministral 8B 2512 on Mistral AI (0% less). Savings are on a mix of 3 input tokens to 1 output token.

When was Qwen 3 32B released?

Alibaba Qwen released it on 28 Apr 2025.

What is Qwen 3 32B's API model id?

Use qwen-3-32b with Vercel AI Gateway.

Sources

Raw catalog entry for qwen-3-32b on Vercel AI Gateway
{
  "input_cost_per_token": 1e-7,
  "litellm_provider": "vercel_ai_gateway",
  "max_input_tokens": 40960,
  "max_output_tokens": 16384,
  "mode": "chat",
  "output_cost_per_token": 3e-7,
  "supports_function_calling": true,
  "supports_tool_choice": true
}