Qwen 3 32B - Kyma API

Overview

Qwen 3 32B is a text-only language model built by Alibaba. It is designed around technical workloads, handling programming, mathematics, and translation across multiple languages. The model natively supports tool use, structured outputs, and explicit reasoning steps. On Kyma, it runs with a 32,768-token context window and an 8,192-token maximum output. It operates in the fast and cheap tiers. Prompt caching is fully supported, billing repeated prompt prefixes at this model’s cached input rate. The model is strictly text-in, text-out and does not support vision or multimodal inputs. While it handles reasoning and tool calls reliably, its output limit means it is not suited for generating extremely long documents or extended codebases in a single pass. Kyma routes requests through an OpenAI-compatible endpoint with automatic failover and exact cost reporting.

Specs

Field	Value
Model ID	`qwen-3-32b`
Best for	Code, math, multilingual
Context window	32,768
Max output tokens	8K
Input modalities	Text
Output modalities	Text
Tool calling	Yes
Structured outputs	Yes
Prompt caching	Yes
Speed	Fast
Cost band	Cheap
Release stage	Stable

Pricing

	Per 1M tokens
Input	$0.11
Output	$0.38

Use this when

Automated Code Generation — Generates and refines code snippets across multiple programming languages with tool-use support.
Mathematical Problem Solving — Breaks down complex math problems using explicit reasoning steps to reach accurate solutions.
Multilingual Text Translation — Translates technical and conversational text across supported languages while preserving context.
Interactive Chat Assistants — Powers responsive conversational agents that leverage structured outputs and automatic failover.

Not ideal for

Tasks requiring image or video understanding, or workloads that need outputs exceeding 8,192 tokens in a single response.

Pick something else when

You need vision or multimodal input: use gemini-3.5-flash-lite or gemini-3.6-flash.
You need longer output sequences: use qwen-3.7-max or deepseek-v3.

Example

from openai import OpenAI

client = OpenAI(base_url="https://kymaapi.com/v1", api_key="ky-...")

response = client.chat.completions.create(
    model="qwen-3-32b",
    messages=[{"role": "user", "content": "..."}],
    tools=[...],  # function calling supported
)

Agent query example

Ask the API which models fit, instead of hardcoding an id:

curl "https://kymaapi.com/v1/models?recommended_for=chat&tools=true&quality_tier=strong"

FAQ

Does this model support function calling and structured JSON outputs? Yes, it natively supports tool use and structured outputs, making it compatible with standard OpenAI-compatible SDKs on Kyma. How does Kyma handle request routing for this model? Every request uses automatic failover; if a serving path degrades, Kyma reroutes the call transparently while reporting the exact model used in the X-Kyma-Model header. Can I reduce costs for repetitive prompts? Yes, prompt caching is enabled, billing repeated prompt prefixes at this model’s cached input rate.

​Overview

​Specs

​Pricing

​Use this when

​Not ideal for

​Pick something else when

​Example

​Agent query example

​FAQ