We introduce our first-generation reasoning models, DeepSeek-R1-Zero and DeepSeek-R1. DeepSeek-R1-Zero is trained via large-scale reinforcement learning (RL) without supervised fine-tuning (SFT) as an initial stage, and it delivers outstanding reasoning capabilities. Through RL training, DeepSeek-R1-Zero naturally exhibits numerous powerful and intriguing reasoning behaviors.
Nevertheless, DeepSeek-R1-Zero suffers from issues such as endless repetition, poor readability, and mixed-language outputs. To address these flaws and further boost reasoning performance, we developed DeepSeek-R1, which incorporates cold-start data prior to the RL phase. DeepSeek-R1 achieves performance comparable to OpenAI o1 on mathematical, coding, and reasoning tasks.
To support the research community, we have open-sourced DeepSeek-R1-Zero, DeepSeek-R1, as well as six dense models distilled from DeepSeek-R1 based on the Llama and Qwen architectures. DeepSeek-R1-Distill-Qwen-32B outperforms OpenAI o1-mini across a wide range of benchmarks, setting a new state-of-the-art record among dense models.
Integrated Deployment
Out-of-the-box inference scripts with pre-configured hardware and software parameters
Released FlagOS-Iluvatar container image supporting deployment within minutes
Consistency Validation
Rigorously evaluated through benchmark testing: Performance and results from the FlagOS software stack are compared against native stacks on multiple public.
Evaluation Results
Benchmark Result
Metrics
DeepSeek-R1-Distill-Qwen-1.5B-Nvidia-Origin
DeepSeek-R1-Distill-Qwen-1.5B-Iluvatar-FlagOS
musr_generative
0.3320
0.3475
mmlu_pro
0.1747
0.1781
aime
0
0
livebench_new
0.1261
0.1229
gpqa_generative_cot
0.0866
0.0914
User Guide
Environment Setup
Item
Version
Docker Version
Docker version 20.10.25, build 20.10.25-0ubuntu1~20.04.1
Choose the appropriate version for your OS (Windows/macOS/Linux)
Follow the installation wizard to complete the setup
2. Configuration
Launch AnythingLLM
Open settings (bottom left, fourth tab)
Configure core LLM parameters
Click "Save Settings" to apply changes
3. Model Interaction
After model loading is complete:
Click "New Conversation"
Enter your question (e.g., “Explain the basics of quantum computing”)
Click the send button to get a response
Technical Overview
FlagOS is a fully open-source system software stack designed to unify the "model–system–chip" layers and foster an open, collaborative ecosystem. It enables a “develop once, run anywhere” workflow across diverse AI accelerators, unlocking hardware performance, eliminating fragmentation among vendor-specific software stacks, and substantially lowering the cost of porting and maintaining AI workloads. With core technologies such as the FlagScale, together with vllm-plugin-fl, distributed training/inference framework, FlagGems universal operator library, FlagCX communication library, and FlagTree unified compiler, the FlagRelease platform leverages the FlagOS stack to automatically produce and release various combinations of <chip + open-source model>. This enables efficient and automated model migration across diverse chips, opening a new chapter for large model deployment and application.
FlagGems
FlagGems is a high-performance, generic operator libraryimplemented in Triton language. It is built on a collection of backend-neutralkernels that aims to accelerate LLM (Large-Language Models) training and inference across diverse hardware platforms.
FlagTree
FlagTree is an open source, unified compiler for multipleAI chips project dedicated to developing a diverse ecosystem of AI chip compilers and related tooling platforms, thereby fostering and strengthening the upstream and downstream Triton ecosystem. Currently in its initial phase, the project aims to maintain compatibility with existing adaptation solutions while unifying the codebase to rapidly implement single-repository multi-backend support. Forupstream model users, it provides unified compilation capabilities across multiple backends; for downstream chip manufacturers, it offers examples of Triton ecosystem integration.
FlagScale and vllm-plugin-fl
Flagscale is a comprehensive toolkit designed to supportthe entire lifecycle of large models. It builds on the strengths of several prominent open-source projects, including Megatron-LM and vLLM, to provide a robust, end-to-end solution for managing and scaling large models.
vllm-plugin-fl is a vLLM plugin built on the FlagOS unified multi-chip backend, to help flagscale support multi-chip on vllm framework.
FlagCX
FlagCX is a scalable and adaptive cross-chip communication library. It serves as a platform where developers, researchers, and AI engineers can collaborate on various projects, contribute to the development of cutting-edge AI solutions, and share their work with the global community.
FlagEval Evaluation Framework
FlagEval is a comprehensive evaluation system and open platform for large models launched in 2023. It aims to establish scientific, fair, and open benchmarks, methodologies, and tools to help researchers assess model and training algorithm performance. It features:
Multi-dimensional Evaluation: Supports 800+ modelevaluations across NLP, CV, Audio, and Multimodal fields,covering 20+ downstream tasks including language understanding and image-text generation.
Industry-Grade Use Cases: Has completed horizonta1 evaluations of mainstream large models, providing authoritative benchmarks for chip-model performance validation.
