Sold by: Engineering
Deployed on AWS
EngGPT 2-16B-A3B e' il Large Language Model italiano sviluppato da Engineering Group per ottimizzare i processi di business garantendo controllo, efficienza e trasparenza nell'adozione dell'AI, in conformita' all'AI ACT.
Overview
EngGPT 2-16B-A3B e' il Large Language Model italiano sviluppato da Engineering Group per ottimizzare i processi di business garantendo controllo, efficienza e trasparenza nell'adozione dell'AI. Sviluppato from scratch in Italia, senza utilizzare modelli preesistenti, garantisce trasparenza sui processi di training e conformita' nativa all'AI Act. Progettato per contesti enterprise e PA, grazie all'architettura Mixture-of-Experts ottimizza costi di training e di inferenza garantendo prestazioni elevate, risultati accurati e ampia possibilita' di personalizzazione sul proprio dominio specifico. E' rilasciato open weight per analisi indipendenti e test, con piena disclosure del processo di training all'interno del technical report. Include la capacita' di doppio ragionamento e risposta del modello: una specifica ed una rapida, in lingue diverse in base all'esigenza. Nativamente ottimizzato per la lingua italiana, si integra nei sistemi esistenti per accelerare casi d'uso come assistenza knowledge-intensive, automazione di workflow complessi, analisi di dati complessi in NL, interpretazione di schemi industriali. E' agent-ready per orchestrare flussi agentic e automazioni su processi e use case mission-critical. Disponibile come servizio su marketplace AWS per un accesso immediato, con standard di sicurezza certificati e scalabilita' globale.
Con l'uso, si accetta il seguente AUP https://www.eng.it/content/dam/eng-portal/resources/documents/legal/Engineering_EngGPT_AUP_2026.03.10.pdf
Highlights
- Open Weight & Auditability - il modello e' pubblicato su HuggingFace, con pesi, documentazione e benchmark pubblici. Anche il processo di training e' trasparente: dataset, pipeline, tecniche, metriche e relative limitazioni sono disponibili nel "Technical Report" suArxiv.
- Training and Inference efficiency - Architettura Mixture-of-Experts e double reasoning consentono output di qualita' elevata e costi sotto controllo. Agent-ready per orchestrare e ottimizzare i processi di business.
- Customizability - Costi di training ridotti grazie al MoE consentono una rapida ed efficace verticalizzazione sul proprio dominio. Ottimizzato per la lingua e il contesto italiano, e' disponibile nelle 6 principali lingue europee (Italiano, Inglese, Tedesco, Spagnolo, Portoghese e Francese).
Details
Sold by
Delivery method
Latest version
Deployed on AWS
New
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
AWS Marketplace now accepts line of credit payments through the PNC Vendor Finance program. This program is available to select AWS customers in the US, excluding NV, NC, ND, TN, & VT.
Pricing
Pricing is based on actual usage, with charges varying according to how much you consume. Subscriptions have no end date and may be canceled any time.
Additional AWS infrastructure costs may apply. Use the AWS Pricing Calculator to estimate your infrastructure costs.
Dimension | Description | Cost/host/hour |
|---|---|---|
ml.g5.12xlarge Inference (Batch) Recommended | Model inference on the ml.g5.12xlarge instance type, batch mode | $4.00 |
ml.g6e.xlarge Inference (Real-Time) Recommended | Model inference on the ml.g6e.xlarge instance type, real-time mode | $4.00 |
ml.g4dn.12xlarge Inference (Batch) | Model inference on the ml.g4dn.12xlarge instance type, batch mode | $4.00 |
ml.g4dn.12xlarge Inference (Real-Time) | Model inference on the ml.g4dn.12xlarge instance type, real-time mode | $4.00 |
ml.g5.12xlarge Inference (Real-Time) | Model inference on the ml.g5.12xlarge instance type, real-time mode | $4.00 |
ml.g5.24xlarge Inference (Batch) | Model inference on the ml.g5.24xlarge instance type, batch mode | $4.00 |
ml.g5.24xlarge Inference (Real-Time) | Model inference on the ml.g5.24xlarge instance type, real-time mode | $4.00 |
ml.g5.48xlarge Inference (Batch) | Model inference on the ml.g5.48xlarge instance type, batch mode | $4.00 |
ml.g5.48xlarge Inference (Real-Time) | Model inference on the ml.g5.48xlarge instance type, real-time mode | $4.00 |
ml.g6e.12xlarge Inference (Real-Time) | Model inference on the ml.g6e.12xlarge instance type, real-time mode | $4.00 |
Dimensions summary
You pay by the hour for each host running model inference, so cost scales with instance size and how long you run it. Pricing splits across three instance families: ml.g4dn.12xlarge, several ml.g5 sizes, and a range of ml.g6e sizes from xlarge up to 48xlarge. The ml.g4dn.12xlarge and ml.g5 instances each offer two modes: batch, for grouped jobs, and real-time, for live requests. The ml.g6e instances are real-time only. Larger instances carry higher hourly rates. You choose the family, size, and mode that fit your workload.
Top-of-mind questions for buyers
Both meter running host-hours, but they serve different workload types. Batch mode groups many requests into scheduled jobs, suited to bulk processing. Real-time mode keeps a host live to answer requests as they arrive. You choose the mode that matches your workload, and you pay for each running host by the hour.
Charges accrue per running host-hour. A host that is fully stopped does not accrue software charges. If you keep a real-time host live to answer requests, it bills for the whole time it runs, whether or not requests arrive. Underlying AWS infrastructure fees may still apply separately.
One host-hour is one hour of one running instance hosting the model for inference. The model uses a Mixture-of-Experts design with about 16 billion total parameters and 3 billion active per token. You pick an instance family and size, and each running host bills by the hour.
www.eng.it
Helpful?
Vendor refund policy
No refund allowed
Legal
Vendor terms and conditions
Upon subscribing to this product, you must acknowledge and agree to the terms and conditions outlined in the vendor's End User License Agreement (EULA) .
Content disclaimer
Vendors are responsible for their product descriptions and other product content. AWS does not warrant that vendors' product descriptions or other product content are accurate, complete, reliable, current, or error-free.
Delivery details
Amazon SageMaker model
An Amazon SageMaker model package is a pre-trained machine learning model ready to use without additional training. Use the model package to create a model on Amazon SageMaker for real-time inference or batch processing. Amazon SageMaker is a fully managed platform for building, training, and deploying machine learning models at scale.
Deploy the model on Amazon SageMaker AI using the following options:
Version release notes
- Primo rilascio del modello LLM deployato su AWS SageMaker AI
- Supporto inferenza real-time e batch, scalabile
- Pronto per integrazione pipeline e applicazioni cloud-native
Additional details
Inputs
- Summary
{ "model":"EngGPTMoe", "messages": [ { "role": "user", "content": "" } ], "chat_template_kwargs": { "enable_thinking": true | false , "reasoning_lang":"ita" | "en", "enable_turbo": true | false } }
{
"model":"EngGPTMoe",
"messages": [
{
"role": "user",
"content": "who are you?"
}
],
"chat_template_kwargs": {
"enable_thinking": false ,
"reasoning_lang": "en",
"enable_turbo": true
}
}
{"messages": [{"role": "user","content": "who are you?"}], "chat_template_kwargs": { "enable_thinking": false ,"reasoning_lang": "en","enable_turbo": true }}
Input data descriptions
The following table describes supported input data fields for real-time inference and batch transform.
Field name | Description | Constraints | Required |
|---|---|---|---|
model | The model identifier to use. See /v1/models for available models. | String | Yes |
messages | List of message objects forming the conversation. | Array of strings | Yes |
temperature | Sampling temperature between 0.0 and 2.0. Lower values make output more deterministic. | float, between 0.0 and 2.0 | No |
max_tokens | Maximum number of tokens to generate in the response. | integer | No |
stream | If true, responses are streamed as Server-Sent Events (SSE). | Boolean (true, false) | No |
stop | Up to 4 sequences at which the model will stop generating further tokens. | string or array of strings | Yes |
Support
Vendor support
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.
Customer reviews
No customer reviews yet
Be the first to review this product . We've partnered with PeerSpot to gather customer feedback. You can share your experience by writing or recording a review, or scheduling a call with a PeerSpot analyst.