InclusionAI/Ring-2.6-1T is now open-sourced
The inclusionAI team has open-sourced Ring-2.6-1T, a trillion-parameter reasoning model designed for complex, real-world task execution. The model supports advanced agent capabilities, adaptive reasoning intensity, and a novel asynchronous reinforcement learning framework. It is available for use via Hugging Face, Docker, SGLang, vLLM, and other deployment methods.
- ▪Ring-2.6-1T is a trillion-parameter AI model focused on real-world complex task execution.
- ▪It features enhanced agent capabilities, allowing it to plan, use tools, and execute multi-step workflows.
- ▪The model supports high and xHigh reasoning effort levels for flexible performance tuning.
- ▪It uses an innovative asynchronous reinforcement learning training method called Async RL with IcePop algorithm.
- ▪Ring-2.6-1T is available under the MIT license and can be deployed using Transformers, vLLM, SGLang, and Docker.
Opening excerpt (first ~120 words) tap to expand
inclusionAI / Ring-2.6-1T like 48 Follow inclusionAI 2.08k Text Generation Transformers Safetensors bailing_hybrid conversational custom_code compressed-tensors License: mit Model card Files Files and versions xet Community 1 Deploy Use this model Instructions to use inclusionAI/Ring-2.6-1T with libraries, inference providers, notebooks, and local apps. Follow these links to get started. Libraries Transformers How to use inclusionAI/Ring-2.6-1T with Transformers: # Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="inclusionAI/Ring-2.6-1T", trust_remote_code=True) messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages) # Load model directly from transformers import AutoModelForCausalLM model =…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Huggingface.
Discussion
0 commentsMore from Huggingface
