Hypnos i1-8B is a professional reasoning model based on Nous Hermes 3 (Llama 3.1 8B), focusing on complex logic, chain-of-thought reasoning, and mathematical problem-solving. This model has conducted unique exploration in the field of hybrid quantum-classical machine learning, and improved the model's creativity and reasoning ability by introducing quantum noise injection.

Natural Language ProcessingTransformersEnglish

Olmo 3 7B RL-Zero Math is a 7-billion-parameter language model developed by Allen AI, specifically optimized for mathematical reasoning tasks. It is trained on mathematical datasets using the RL-Zero reinforcement learning method, which can effectively improve mathematical reasoning ability.

Natural Language ProcessingTransformersEnglish

metatune-gpt20b is a large language model prototype with self-improvement ability. It can generate new data for itself, evaluate its own performance, and adjust hyperparameters according to improvement indicators. This model performs excellently in scientific and mathematical understanding at the postdoctoral level and can also be used for coding tasks.

Natural Language ProcessingTransformersEnglish

This project distills the inference ability of GPT into the Llama-3.1-8B model through an innovative two-stage training process. First, knowledge distillation and format alignment are performed through supervised fine-tuning, and then reinforcement learning is used to encourage the model to independently explore and optimize inference strategies, focusing on breakthroughs in the field of mathematical reasoning.

Natural Language ProcessingTransformersMultiple Languages

QuestA is an innovative framework that enhances the reasoning ability of large language models through a question-enhanced approach. It integrates partial solutions during the reinforcement learning training process, significantly improving the model's performance on complex tasks such as mathematical reasoning, especially achieving optimal results on small-parameter models.

Natural Language ProcessingTransformers

MobileLLM-R1 is a series of efficient reasoning models launched by Meta, focusing on solving mathematical, programming, and scientific problems. The model offers three scale versions of 140M, 360M, and 950M, with excellent reasoning ability and reproducibility.

Natural Language ProcessingTransformersEnglish

OpenReasoning-Nemotron-32B is a large language model based on Qwen2.5-32B-Instruct, which has been post-trained specifically for reasoning tasks in generating mathematical, code, and scientific solutions. This model performs excellently in various reasoning benchmark tests, with high accuracy and strong reasoning ability, and supports an output of up to 64K tokens.

Natural Language ProcessingTransformersMultiple Languages

A LoRA fine-tuned version based on Qwen3-1.7B, which enhances the reasoning ability for complex logic and mathematical tasks through adaptive reasoning and thinking program techniques, supporting reasoning methods such as chain-of-thought and leap-of-thought.

Natural Language ProcessingSafetensorsOther

This model is a fine-tuned version based on DeepSeek-R1-0528-Qwen3-8B, trained using TRL and GRPO methods, focusing on improving mathematical reasoning ability.

Natural Language ProcessingTransformersMultiple Languages

A language model fine-tuned based on Qwen3-1.7B, which improves mathematical reasoning ability through reinforcement learning technology

Natural Language ProcessingSafetensorsEnglish

A fine-tuned version based on Qwen3-1.7B, which enhances mathematical reasoning ability through one-shot reinforcement learning and verifiable reward (RLVR) methods, and performs excellently in mathematical benchmark tests and coding tasks.

Natural Language ProcessingSafetensors

J-1 is a reasoning model based on the Gemma 3 4-B model, trained on Italian mathematical and reasoning problems using the GRPO method. On the Invalsi ITA test dataset, the performance has been improved from 57% of the base model to 79%, significantly enhancing the Italian reasoning ability.

Natural Language ProcessingTransformersOther

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

Fathom-R1-14B is a project based on the R1-distilled-14B model, achieving o4-mini level mathematical reasoning ability under a 16K context with a low training cost of $499.

Natural Language ProcessingTransformers

GLM-4-Z1-Rumination-32B-0414 is an inference model with in-depth reflective ability in the GLM series. It has 32 billion parameters and performs excellently in mathematical reasoning and complex task solving. It supports multiple function call functions.

Natural Language ProcessingTransformersMultiple Languages

Qwen2.5-14B-it-restore is a model generated by combining the della, ties, and model stock fusion methods based on the Qwen2.5-14B and Qwen2.5-14B-Instruct models, aiming to address the decline in instruction-following capability and weakened mathematical ability.

Natural Language ProcessingSafetensorsMultiple Languages

MathCoder2 is a model created through continuous pre - training on the MathCode - Pile mathematical code dataset. It is specifically designed to provide strong support for advanced mathematical reasoning tasks and has significantly improved mathematical reasoning ability.

Natural Language ProcessingSafetensorsEnglish

Qwen2.5-72B-Instruct is a large-scale language model developed by the Tongyi Qianwen team at Alibaba, with 72.7 billion parameters. This model has significant improvements in knowledge understanding, coding ability, mathematical reasoning, and multilingual support. It supports a context length of up to 128K tokens and can generate content of up to 8K tokens.

Natural Language ProcessingGgufMultiple Languages

MetaMath-Mistral-7B is a mathematical reasoning model fine-tuned on the MetaMathQA dataset based on the Mistral-7B model, which significantly improves the ability to solve mathematical problems.

Natural Language ProcessingTransformers

This model is a LaTeX OCR model fine-tuned based on Nougat-base, specifically designed to generate LaTeX code from images, with a particular optimization for the recognition ability of mathematical formula images.

MultimodalTransformersEnglish