DeepSeek-GRM-16B is a generative reward model based on Self-Principled Critique Tuning (SPCT), which can generate a transparent 'Principle → Critique → Score' evaluation process for query-responses and can be used in tasks such as reinforcement learning, evaluation, and data collection of large language models.

Natural Language ProcessingTransformersMultiple Languages

DeepSeek-V3.1 is a large language model that is post-trained based on DeepSeek-V3.1-Base. It has a 128K context length and supports mixed thinking modes, intelligent tool invocation, and code proxy functions. This model performs excellently in multiple benchmark tests, especially showing significant improvements in mathematical reasoning, code generation, and search agent tasks.

Natural Language ProcessingTransformersMultiple Languages

This is a quantized version of the DeepSeek-V3.1 model, which uses the imatrix technology of llama.cpp for quantization processing. The aim is to improve the operating efficiency and performance of the model on different hardware. This version offers multiple quantization levels, ranging from high-quality Q8_0 to extremely low-quality IQ1_M, to meet the requirements of different hardware configurations and performance needs.

Natural Language ProcessingGguf

DeepSeek-V3.1 is a large language model developed by DeepSeek-AI and is an upgraded version of DeepSeek-V3. This model supports both hybrid thinking mode and non-thinking mode, and performs excellently in tool invocation, code generation, mathematical reasoning, etc., supporting a context length of 128K.

Natural Language ProcessingTransformers

DeepSeek-R1 is an inference model trained through large-scale reinforcement learning. It performs excellently in mathematics, code, and reasoning tasks, and can demonstrate powerful reasoning abilities without supervised fine-tuning, including self-verification, reflection, and generating long thought chains, etc.

Natural Language ProcessingTransformersEnglish

AceReason-Nemotron-7B is a mathematical and code reasoning model trained through reinforcement learning. It is developed based on DeepSeek-R1-Distilled-Qwen-7B and performs excellently in multiple reasoning benchmark tests.

Natural Language ProcessingTransformersMultiple Languages

This is a language model specifically optimized for the Indonesian legal field. It is fine-tuned on the Indonesian legal Q&A dataset using the GRPO method based on DeepSeek-R1-0528-Qwen3-8B, focusing on improving legal reasoning and structured thinking abilities.

Natural Language ProcessingTransformers

AceReason-Nemotron-7B is a mathematical and code reasoning model trained based on reinforcement learning. It starts training from DeepSeek-R1-Distilled-Qwen-7B and performs excellently in multiple benchmark tests.

Natural Language ProcessingTransformers

DeepSeek-R1-0528 is an upgraded large language model launched by DeepSeek. It has significant improvements in reasoning ability and reducing hallucination rate, and its overall performance is close to that of leading models.

Natural Language ProcessingTransformers

A math and code reasoning model trained through reinforcement learning, based on DeepSeek-R1-Distilled-Qwen-7B, excelling in mathematical and code reasoning tasks

Natural Language ProcessingTransformersMultiple Languages

AceReason-Nemotron-14B is a math and code reasoning model trained through reinforcement learning, based on DeepSeek-R1-Distilled-Qwen-14B, excelling in math and code reasoning tasks.

Natural Language ProcessingTransformersMultiple Languages

AceMath-RL-Nemotron-7B is a mathematical reasoning model trained entirely through reinforcement learning. It is trained based on Deepseek-R1-Distilled-Qwen-7B and performs excellently in mathematical reasoning tasks. It also has certain generalization ability in coding tasks.

Natural Language ProcessingTransformersEnglish

A code-reasoning large language model fine-tuned based on DeepSeek-R1-Distilled-Qwen-1.5B, utilizing distributed reinforcement learning technology to extend long-context processing capabilities

Natural Language ProcessingGgufEnglish

DeepCoder-1.5B-Preview is a large language model for code reasoning, fine-tuned from DeepSeek-R1-Distilled-Qwen-1.5B through distributed reinforcement learning, capable of handling longer context lengths.

Natural Language ProcessingTransformersEnglish

A code reasoning large model fine-tuned based on DeepSeek-R1-Distilled-Qwen-1.5B, utilizing distributed reinforcement learning technology to enhance long-context processing capabilities

Natural Language ProcessingTransformersEnglish

The AWQ quantized version of DeepSeek V3 0324, quantized by Eric Hartford and v2ray. This version fixes the overflow issue that occurs when using float16, optimizes the model performance, and supports running at the full context length on 8 80GB GPUs.

Natural Language ProcessingTransformersMultiple Languages

Light-R1-32B-DS is a near-SOTA level 32B mathematical model, fine-tuned based on DeepSeek-R1-Distill-Qwen-32B, achieving high performance with only 3K SFT data.

Natural Language ProcessingTransformers

LeX-Enhancer is a lightweight prompt enhancement model distilled from DeepSeek-R1, designed to generate high-quality, visually rich prompt outputs suitable for text-to-image generation.

Natural Language ProcessingTransformersOther

This is the GGUF quantized version of the DeepSeek-V3-1B-Test model, with quantization services provided by TensorBlock. This model is a small language model with a scale of 1B parameters, offering multiple quantization levels to meet different hardware and performance requirements.

Natural Language ProcessingTransformers

ARWKV-R1-1B5 is an early preview version of a 7-billion-parameter model based on RNN, trained through three-stage knowledge distillation from DeepSeek-R1-Distill-Qwen-1.5B, with a context length of 2k.

Natural Language ProcessingTransformersMultiple Languages