gpt‑oss‑20b is an open-weight reasoning model with high performance and low latency. Its configurable reasoning level (low, medium, high) adjusts analysis depth to task criticality. It supports function calling, structured JSON output and streaming for automated workflows and intelligent agents. Its Mixture of Experts architecture activates 3.6 billion of 21 billion parameters, delivering computational efficiency without sacrificing response quality.
Overview
Technical specifications
- PROVIDER LOCATION: Non-EU
- PARAMETERS: 21.5B (3.6B active)
- MAX CONTEXT: 131K token
- LICENSE: Apache 2.0
- FEATURES: Function calling, Reasoning
- OUTPUT FORMATS: Raw_text, Json_object, Json_schema
Use cases
- Advanced document analysis
- Intelligent agents with function calling
- Automated classification and triage
- Structured data extraction
Implementation examples
Input
curl --request POST \
--url https://apiaruba.it/aiplatform/v1/chat/completions \
--header 'content-type: application/json' \
--header 'x-aruba-api-key: *****' \
--data '{
"model": "gpt-oss-20b",
"messages": [
{
"role": "system",
"content": "Sei un assistente conciso."
},
{
"role": "user",
"content": "Spiega cos'\''è un embedding."
}
],
"temperature": 0.7,
"max_tokens": 1000
}'
Output
{
"id": "chatcmpl-9abcc998-9e21-4fca-b610-f280e96b0f79",
"created": 1787560923,
"model": "gpt-oss-20b",
"object": "chat.completion",
"system_fingerprint": "vllm-0.24.0-c6d97d58",
"choices": [
{
"finish_reason": "stop",
"index": 0,
"message": {
"content": "Un **embedding** è una rappresentazione numerica (vettore) di un oggetto (parola, frase, immagine, nodo, ecc.) in uno spazio continuo. \n- Permette di trasformare dati discreti o categorici in numeri che i modelli di machine‑learning possono elaborare. \n- Gli embedding cercano di mantenere relazioni semantiche: elementi simili ottengono vettori vicini. \n- Esempi: Word2Vec/GloVe per parole, BERT per frasi, embeddings grafici per nodi di rete.",
"role": "assistant",
"reasoning_content": "We need to produce concise answer in Italian. The user says: \"Spiega cos'è un embedding.\" We need to explain what an embedding is. They want concise. So something like: \"Un embedding è una rappresentazione vettoriale di un elemento (parole, immagini, ecc.) in uno spazio continuo per permettere algoritmi di machine learning di operare su di esso.\" Provide definition, purpose, examples: word embeddings like Word2Vec, GloVe, sentence embeddings, etc. Keep concise.\n\nLet's answer in Italian concisely.",
"provider_specific_fields": {
"reasoning": "We need to produce concise answer in Italian. The user says: \"Spiega cos'è un embedding.\" We need to explain what an embedding is. They want concise. So something like: \"Un embedding è una rappresentazione vettoriale di un elemento (parole, immagini, ecc.) in uno spazio continuo per permettere algoritmi di machine learning di operare su di esso.\" Provide definition, purpose, examples: word embeddings like Word2Vec, GloVe, sentence embeddings, etc. Keep concise.\n\nLet's answer in Italian concisely.",
"refusal": null
}
},
"provider_specific_fields": {}
}
],
"usage": {
"completion_tokens": 245,
"prompt_tokens": 89,
"total_tokens": 334
}
}
Alternatives in the catalog
Cost
0,0237 €
+ VAT / 1M token (input)
0,1146 €
+ VAT / 1M token (output)