alibaba/qwen3-32bOpen WeightsDense open Qwen model for self-hosted chat, reasoning, and coding
Benchmark Pricing$0.70 / 1M in$2.80 / 1M out
Context: 131,072 tokens•Max Output: 16,384 tokens•Knowledge: 2025-04•Released: 2025-04-28
TECHNICAL SPECIFICATIONS
Context Window131,072131.1K tokens
Max Output16,38416.4K tokens
LicenseOpen SourceSelf-hostable
Input Modalities
Text
Function / Tool CallingSupported (MCP Ready)
Structured OutputsGuaranteed Schema
MODEL CONTEXT PROTOCOL (MCP) READINESS
✓ Qwen3 32B natively supports OpenAI / Anthropic format tool calling and is fully compatible with MCP servers and agent loops.
MCP Context Window Recommendations
With a context capacity of 131,072 tokens, this model can handle moderately large MCP tool definitions and multi-turn conversation logs comfortably.
STANDARDIZED BENCHMARKS & EVALS
Aider Polyglotpercent correct
40API CALL INVOCATION EXAMPLE
// Example: Calling Qwen3 32B via OpenAI SDK
import OpenAI from "openai";
const client = new OpenAI({
apiKey: process.env.ALIBABA_API_KEY,
});
const response = await client.chat.completions.create({
model: "alibaba/qwen3-32b",
messages: [{ role: "user", content: "Hello! Can you inspect MCP tools?" }],
});
console.log(response.choices[0].message.content);PRICING BY PROVIDER
Alibabaqwen3-32b
$0.70 / $2.80
Cortecsqwen3-32b
$0.18 / $0.70
NanoGPTqwen/qwen3-32b
$0.10 / $0.30$0.05 cache
LLM Gatewayscx-ai/qwen3-32b
$0.36 / $0.87
AbacusQwen/Qwen3-32B
$0.08 / $0.28
Merge Gatewayqwen/qwen3-32b
$0.15 / $0.60
Hugging FaceQwen/Qwen3-32B
$0.29 / $0.59
Amazon Bedrockqwen.qwen3-32b-v1:0
$0.15 / $0.60
Kilo Gatewayqwen/qwen3-32b
$0.08 / $0.28
DevPass (LLM Gateway)qwen3-32b
$0.36 / $0.87
PioneerQwen/Qwen3-32B
$0.90 / $0.90$0.90 cache
Alibaba (China)qwen3-32b
$0.29 / $1.15
Deep InfraQwen/Qwen3-32B
$0.08 / $0.28
iFlowqwen3-32b
$0.00 / $0.00
Jiekou.AIqwen/qwen3-32b-fp8
$0.10 / $0.45
Rates in USD per 1M tokens. Rates synchronized from models.dev public catalog.
MODEL RESOURCES
More Models by Alibaba
Alternative and family models from the same lab
Qwen3 Coder Plus
alibaba/qwen3-coder-plus•Alibaba
Hosted Qwen coder for software agents, repo edits, and long-context code
65.5K out max
In: $1.00 • Out: $5.00
AlibabaQwen Turbo
alibaba/qwen-turbo•Alibaba
Efficient Qwen model for fast chat, extraction, and high-volume workloads
16.4K out max
In: $0.05 • Out: $0.20
AlibabaQwen3.8 Max
alibaba/qwen3.8-max•Alibaba
2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows
131.1K out maxVisionPDFTerminal-Bench: 86.6 avg@10
In: $2.00 • Out: $6.00
Alibaba