# Nebius Token Factory model catalog

Use this catalog to compare public models available in Nebius Token Factory. It includes model names, model IDs, providers, regions, context windows and pricing values for inference, embeddings, reranking and image generation.

For integrations and automation, fetch `/api/public/models_info`. The JSON endpoint is the authoritative machine-readable source; this Markdown view is optimized for reading and quick comparison.

Pricing values map to these JSON fields:

- Token models: `flavors[].input_price_per_million_tokens` and `flavors[].output_price_per_million_tokens`.
- Embedding and rerank models: `flavors[].input_price_per_million_tokens`.
- Image models: `flavors[].price_per_image` and optional `flavors[].price_per_image_with_lora`.

| Model | Vendor | Type | Parameters, B | Context window, K tokens | Regions | Pricing |
| --- | --- | --- | ---: | ---: | --- | --- |
| DeepSeek-V4-Pro | deepseek | text2text | 862 | 1000 | uk-south1 | deepseek-ai/DeepSeek-V4-Pro: input\_price\_per\_million\_tokens=1.75, output\_price\_per\_million\_tokens=3.5 |
| Gemma-3-27b-it | google | text2text | 27 | 110 | eu-north1 | google/gemma-3-27b-it: input\_price\_per\_million\_tokens=0.1, output\_price\_per\_million\_tokens=0.3 |
| Llama-3.3-70B-Instruct | meta | text2text | 70.6 | 128 | eu-north1 | meta-llama/Llama-3.3-70B-Instruct: input\_price\_per\_million\_tokens=0.13, output\_price\_per\_million\_tokens=0.4 |
| MiniMax-M2.5 | MiniMaxAI | text2text | 229 | 196 | us-central1 | MiniMaxAI/MiniMax-M2.5: input\_price\_per\_million\_tokens=0.3, output\_price\_per\_million\_tokens=1.2 |
| Kimi-K2.6 | moonshotai | image2text | 1026 | 256 | us-central1 | moonshotai/Kimi-K2.6: input\_price\_per\_million\_tokens=0.95, output\_price\_per\_million\_tokens=4 |
| Kimi-K2.7-Code | moonshotai | text2text | 1026 | 256 | us-central1 | moonshotai/Kimi-K2.7-Code: input\_price\_per\_million\_tokens=0.95, output\_price\_per\_million\_tokens=4 |
| Hermes-4-405B | NousResearch | text2text | 405 | 128 | eu-north1 | NousResearch/Hermes-4-405B: input\_price\_per\_million\_tokens=1, output\_price\_per\_million\_tokens=3 |
| Hermes-4-70B | NousResearch | text2text | 70 | 128 | eu-north1 | NousResearch/Hermes-4-70B: input\_price\_per\_million\_tokens=0.13, output\_price\_per\_million\_tokens=0.4 |
| Cosmos3-Super-Reasoner | nvidia | image2text | 33 | 256 | eu-north1 | nvidia/Cosmos3-Super-Reasoner: input\_price\_per\_million\_tokens=0.1, output\_price\_per\_million\_tokens=0.3 |
| Nemotron-3-Nano-30B-A3B | nvidia | text2text | 30 | 262 | eu-north1 | nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B: input\_price\_per\_million\_tokens=0.06, output\_price\_per\_million\_tokens=0.24 |
| Nemotron-3-Super-120b-a12b | nvidia | text2text | 120 | 256 | us-central1 | nvidia/nemotron-3-super-120b-a12b: input\_price\_per\_million\_tokens=0.3, output\_price\_per\_million\_tokens=0.9 |
| Nemotron-3-Ultra-550b-a55b | nvidia | text2text | 550 | 1000 | us-central1 | nvidia/Nemotron-3-Ultra-550b-a55b: input\_price\_per\_million\_tokens=1, output\_price\_per\_million\_tokens=3 |
| gpt-oss-120b | openai | text2text | 120 | 131 | eu-north1 | openai/gpt-oss-120b: input\_price\_per\_million\_tokens=0.15, output\_price\_per\_million\_tokens=0.6 |
| openbmb/MiniCPM-V-4\_5 | openbmb | image2text | 8 | 32 | eu-north1 | openbmb/MiniCPM-V-4\_5: input\_price\_per\_million\_tokens=0.658, output\_price\_per\_million\_tokens=1.11 |
| Qwen2.5-VL-72B-Instruct | Qwen | image2text | 72 | 32 | eu-north1 | Qwen/Qwen2.5-VL-72B-Instruct: input\_price\_per\_million\_tokens=0.25, output\_price\_per\_million\_tokens=0.75 |
| Qwen3-235B-A22B-Instruct-2507 | Qwen | text2text | 235 | 262 | eu-north1 | Qwen/Qwen3-235B-A22B-Instruct-2507: input\_price\_per\_million\_tokens=0.2, output\_price\_per\_million\_tokens=0.6 |
| Qwen3-30B-A3B-Instruct-2507 | Qwen | text2text | 30.5 | 262 | eu-north1 | Qwen/Qwen3-30B-A3B-Instruct-2507: input\_price\_per\_million\_tokens=0.1, output\_price\_per\_million\_tokens=0.3 |
| Qwen3-32B | Qwen | text2text | 32.8 | 41 | eu-north1 | Qwen/Qwen3-32B: input\_price\_per\_million\_tokens=0.1, output\_price\_per\_million\_tokens=0.3 |
| Qwen3-Embedding-8B | Qwen | embedding | 8 | 41 | eu-north1 | Qwen/Qwen3-Embedding-8B: input\_price\_per\_million\_tokens=0.01, output\_price\_per\_million\_tokens=0 |
| Qwen3-Next-80B-A3B-Thinking | Qwen | text2text | 81 | 128 | eu-north1 | Qwen/Qwen3-Next-80B-A3B-Thinking: input\_price\_per\_million\_tokens=0.15, output\_price\_per\_million\_tokens=1.2 |
| Qwen3.5-397B-A17B | Qwen | text2text | 397 | 262 | us-central1 | Qwen/Qwen3.5-397B-A17B: input\_price\_per\_million\_tokens=0.6, output\_price\_per\_million\_tokens=3.6 |
| GLM-5.1 | zai-org | text2text | 750 | 200 | eu-north1 | zai-org/GLM-5.1: input\_price\_per\_million\_tokens=1.4, output\_price\_per\_million\_tokens=4.4 |
| GLM-5.2 | zai-org | text2text | 750 | 1000 | us-central1 | zai-org/GLM-5.2: input\_price\_per\_million\_tokens=1.4, output\_price\_per\_million\_tokens=4.4 |