comfyui-OreX

comfyui-OreX is a set of nodes for using LLM models via a free API.

Currently, the following models are available:

• deepseek-ai/DeepSeek-R1: A versatile model for general language tasks.
• Qwen/QwQ-32B: Designed for high-performance language understanding and generation.
• meta-llama/Llama-3.2-90B-Vision-Instruct: Combines vision and language capabilities for multimodal tasks.
• deepseek-ai/DeepSeek-R1-Distill-Llama-70B: A distilled version of DeepSeek-R1 for efficient performance.
• meta-llama/Llama-3.3-70B-Instruct: An instruction-tuned model for various language tasks.
• deepseek-ai/DeepSeek-R1-Distill-Qwen-32B: Distilled for efficiency while maintaining high performance.
• deepseek-ai/DeepSeek-R1-Distill-Llama-8B: A smaller, efficient version of the DeepSeek-R1 model.
• deepseek-ai/DeepSeek-R1-Distill-Qwen-14B: Balances performance and efficiency for language tasks.
• deepseek-ai/DeepSeek-R1-Distill-Qwen-7B: A compact model suitable for various applications.
• deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B: Lightweight and efficient for quick language tasks.
• microsoft/phi-4: A robust model for general language understanding and generation.
• mistralai/Mistral-Large-Instruct-2411: Optimized for instruction-following tasks.
• neuralmagic/Llama-3.1-Nemotron-70B-Instruct-HF-FP8-dynamic: Efficient and dynamic for various language tasks.
• google/gemma-2-9b-it: A model designed for Italian language tasks.
• nvidia/AceMath-7B-Instruct: Specialized in mathematical problem-solving.
• CohereForAI/aya-expanse-32b: A large model for expansive language understanding.
• Qwen/Qwen2.5-Coder-32B-Instruct: Optimized for coding and instruction-following tasks.
• THUDM/glm-4-9b-chat: Designed for conversational AI applications.
• CohereForAI/c4ai-command-r-plus-08-2024: A command-focused model for precise instructions.
• tiiuae/Falcon3-10B-Instruct: Efficient and instruction-tuned for various tasks.
• NovaSky-AI/Sky-T1-32B: A large model for comprehensive language tasks.
• bespokelabs/Bespoke-Stratos-32B: Designed for high-performance language understanding.
• netease-youdao/Confucius-o1-14B: A model with a focus on Chinese language tasks.
• Qwen/Qwen2.5-1.5B-Instruct: Lightweight and efficient for instruction-following.
• mistralai/Ministral-8B-Instruct-2410: Compact and optimized for instruction tasks.
• openbmb/MiniCPM3-4B: A smaller model for efficient language processing.
• jinaai/ReaderLM-v2: Specialized in reading comprehension tasks.
• ibm-granite/granite-3.1-8b-instruct: Robust and instruction-tuned for various applications.
• microsoft/Phi-3.5-mini-instruct: A mini version optimized for instruction tasks.
• ozone-ai/0x-lite: Lightweight and efficient for quick language tasks.
• mixedbread-ai/mxbai-embed-large-v1: Designed for embedding and language understanding.

The nodes can operate in the following modes:

• txt to image: Generate images from textual descriptions.
• image 2 image: Modify existing images based on textual instructions.

You can get a free API here: https://ai.io.net/ai/api-keys

by StableDif & OreX