Fine-tuning - verified
Nebius Token Factory fine-tuning: status, methods and prices
Offered. 3 offerings recorded: LoRA, Full and RFT. No training price is stated on the pages we checked. Weight export: Yes and Not stated.
Recorded as: Nebius Token Factory (formerly Nebius AI Studio).
Read from docs.tokenfactory.nebius.com. Compare every fine-tuning API
- Status
- Offered
- Offerings
- 3
- Methods
- LoRA, Full, RFT
- Export weights
- Varies
Status
Is Nebius Token Factory fine-tuning still available?
Offered
All 3 offerings
LoRA supervised fine-tuning
- Status
Available
“You can currently run long-context supervised fine-tuning on full model weights or LoRA adapters.”
docs.tokenfactory.nebius.com - Tunable base models
unsloth/gpt-oss-20b-BF16unsloth/gpt-oss-120b-BF16Qwen/Qwen3-235B-A22BQwen/Qwen3-30B-A3BQwen/Qwen3-Coder-480B-A35B-InstructQwen/Qwen3-32BQwen/Qwen3-14BQwen/Qwen3-8BQwen/Qwen3-4BQwen/Qwen3-1.7B
Subset; Qwen2.5 and other variants also listed.
“unsloth/gpt-oss-120b-BF16, ( Model card) · LoRA and Full Parameter fine-tuning”
docs.tokenfactory.nebius.com “Qwen/Qwen3-32B, ( Model card) · LoRA and Full Parameter fine-tuning”
docs.tokenfactory.nebius.com - Training price
Not stated
docs.nebius.com/studio URLs redirect to Token Factory docs. Post-training offers long-context SFT on full weights or LoRA; reinforcement fine-tuning is a limited professional service. The public prices page redirected to a logged-in console and the billing docs list no fine-tuning rates, so training prices are null.
- Tuned-model inference
- Not stated
- Hosting requirement
dedicated endpoints only
“Deployment options currently only include via Dedicated endpoints.”
docs.tokenfactory.nebius.com - Export weights
Yes
“Download the resulting model checkpoints”
docs.tokenfactory.nebius.com - Data limits
- Not stated
Full-parameter supervised fine-tuning
- Status
Available
“You can currently run long-context supervised fine-tuning on full model weights or LoRA adapters.”
docs.tokenfactory.nebius.com - Tunable base models
deepseek-ai/DeepSeek-V3-0324deepseek-ai/DeepSeek-V4-Flashgoogle/gemma-4-E2B-itgoogle/gemma-4-E4B-itgoogle/gemma-4-31B-itplus all LoRA-table models
“deepseek-ai/DeepSeek-V4-Flash, ( Model card) · Full fine-tuning”
docs.tokenfactory.nebius.com “google/gemma-4-31B-it, (Model card) · Full fine-tuning”
docs.tokenfactory.nebius.com - Training price
Not stated
docs.nebius.com/studio URLs redirect to Token Factory docs. Post-training offers long-context SFT on full weights or LoRA; reinforcement fine-tuning is a limited professional service. The public prices page redirected to a logged-in console and the billing docs list no fine-tuning rates, so training prices are null.
- Tuned-model inference
- Not stated
- Hosting requirement
dedicated endpoints only
“Deployment options currently only include via Dedicated endpoints.”
docs.tokenfactory.nebius.com - Export weights
Yes
“Download the resulting model checkpoints”
docs.tokenfactory.nebius.com - Data limits
- Not stated
Reinforcement fine-tuning (limited professional service)
- Status
Limited access / upcoming
“Reinforcement Fine-tuning (limited professional service”
docs.tokenfactory.nebius.com - Tunable base models
- Not stated
- Training price
Not stated
docs.nebius.com/studio URLs redirect to Token Factory docs. Post-training offers long-context SFT on full weights or LoRA; reinforcement fine-tuning is a limited professional service. The public prices page redirected to a logged-in console and the billing docs list no fine-tuning rates, so training prices are null.
- Tuned-model inference
- Not stated
- Hosting requirement
- Not stated
- Export weights
Not stated
- Data limits
- Not stated
- Notes
Listed under 'Upcoming Advanced post-training options'; access by request.
Research notes
docs.nebius.com/studio URLs redirect to Token Factory docs. Post-training offers long-context SFT on full weights or LoRA; reinforcement fine-tuning is a limited professional service. The public prices page redirected to a logged-in console and the billing docs list no fine-tuning rates, so training prices are null.
Plain language
The methods, explained
- SFTSupervised fine-tuning
- You show the model example prompts with the answers you want, and it learns to answer like that. The default method almost everywhere.
- DPODirect preference optimization (preference tuning)
- You give pairs of answers to the same prompt, one preferred and one not, and the model learns to lean toward the preferred kind. Good for tone and style.
- RFTReinforcement fine-tuning
- The model attempts tasks and a grader (your code, or another model) scores each attempt; training pushes it toward higher-scoring answers. Often billed by training hour, plus the grader's tokens.
- Vision SFTVision fine-tuning
- Supervised fine-tuning where the examples include images, so the model gets better at a visual task.
- DistillationDistillation
- A large teacher model writes answers that are used to fine-tune a smaller, cheaper student model. You pay for the teacher's calls as well as the training.
- LoRALoRA (adapter) fine-tuning
- Trains a small set of extra weights, an adapter, on top of the frozen base model. Cheaper and faster than full fine-tuning; the adapter is served alongside the base model.
- FullFull-parameter fine-tuning
- Updates every weight in the model. Costs more than LoRA and produces a complete copy of the model.
- CPTContinued pre-training
- Keeps training the model on a large body of raw, unlabeled text from your domain, so it absorbs vocabulary and facts before any instruction tuning.
FAQ
Does Nebius Token Factory offer fine-tuning?
Offered. 3 offerings: LoRA, Full and RFT.
How much does Nebius Token Factory fine-tuning cost?
No training price is stated on the pages we checked. Not stated for LoRA supervised fine-tuning, Full-parameter supervised fine-tuning and Reinforcement fine-tuning (limited professional service). Prices are in Nebius Token Factory's own units and are not converted.
Can you download a model you fine-tune on Nebius Token Factory?
Yes. The docs say: "Download the resulting model checkpoints"
Sources
- https://docs.tokenfactory.nebius.com/post-training/overview
- https://docs.tokenfactory.nebius.com/other-capabilities/billing-new
- https://docs.tokenfactory.nebius.com/post-training/models
- https://docs.tokenfactory.nebius.com/post-training/how-to-fine-tune
Last verified . This is a directory of what each provider's own pages say, not a quality ranking. Prices are copied in the provider's own unit and never converted, so rows in different units are not comparable.
Newsletter
Know when a provider changes its fine-tuning terms
One email per week. New model releases, deprecations and pricing changes, and what they mean for what you build.