Note
The context length for all fine-tuning models is 8k.
The context length for all fine-tuning models is 8k.
DeepSeek
| Name | Supported fine-tuning type | License |
|---|---|---|
| deepseek-ai/DeepSeekV3-0324 (Model card) | Full fine-tuning | MIT License |
Meta
Nebius AI Studio and the Meta models hosted in the service are built with Meta Llama 3.1 and 3.3. For more information, see the following Acceptable Use Policies:| Name | Supported fine-tuning type | License |
|---|---|---|
| meta-llama/Llama-3.2-1B-Instruct (Model card) | LoRA and full fine-tuning | Llama 3.2 Community License Agreement |
| meta-llama/Llama-3.2-3B-Instruct (Model card) | LoRA and full fine-tuning | Llama 3.2 Community License Agreement |
| meta-llama/Llama-3.1-8B-Instruct (Model card) | LoRA and full fine-tuning | Llama 3.1 Community License Agreement |
| meta-llama/Llama-3.1-70B (Model card) | LoRA and full fine-tuning | Llama 3.1 Community License Agreement |
| meta-llama/Llama-3.3-70B-Instruct (Model card) | LoRA and full fine-tuning | Llama 3.3 Community License Agreement |
Qwen
Nebius AI Studio and the Qwen models hosted in the service are built with Qwen.| Name | Supported fine-tuning type | License |
|---|---|---|
| Qwen/Qwen3-14B (Model card) | LoRA and full fine-tuning | Apache License 2.0 |
| Qwen/Qwen3-32B (Model card) | LoRA and full fine-tuning | Apache License 2.0 |
Base LoRA adapter models available for deployment
You can deploy serverless LoRA adapter models in Nebius AI Studio. To deploy a model, first prepare or fine-tune it by using one of the base models below:| Name | Supported fine-tuning type | License |
|---|---|---|
| meta-llama/Llama-3.1-8B-Instruct (Model card) | LoRA and full fine-tuning | Llama 3.1 Community License Agreement |
| meta-llama/Llama-3.3-70B-Instruct (Model card) | LoRA | Llama 3.3 Community License Agreement |
| Qwen/Qwen3-32B (Model card) | LoRA | Apache License 2.0 |