DeepSeek-R1 models improve reasoning through reinforcement learning and fine-tuning, outperforming major benchmarks.

DeepSeek-R1 models improve reasoning through reinforcement learning and fine-tuning, outperforming major benchmarks.

Llama 3.2 offers advanced multilingual generative language models for diverse commercial and research applications.

Llama 3.2-Vision models by Meta, integrating advanced image reasoning capabilities with text, enhance visual tasks.

Meta Llama 3.1 is a suite of multilingual, large language models (LLMs) available in 8B,70B, and 405B sizes optimized for text.

Llama Guard 3 improves content safety classification with high accuracy and supports multiple languages.

all-mpnet-base-v2 is a sentence-transformers model for efficient semantic sentence representation and clustering.

Transform and enhance conversational AI with Phi-4 for improved language model performance and safety.

Mistral-7B-Instruct-v0.3 is a fine-tuned model with extended vocabulary and function calling.

DeepSeek-R1 models improve reasoning through reinforcement learning and fine-tuning, outperforming major benchmarks.

DeepSeek-R1 models improve reasoning through reinforcement learning and fine-tuning, outperforming major benchmarks.

Llama 3.2 offers advanced multilingual generative language models for diverse commercial and research applications.

Llama 3.2-Vision models by Meta, integrating advanced image reasoning capabilities with text, enhance visual tasks.

Meta Llama 3.1 is a suite of multilingual, large language models (LLMs) available in 8B,70B, and 405B sizes optimized for text.

Llama Guard 3 improves content safety classification with high accuracy and supports multiple languages.

all-mpnet-base-v2 is a sentence-transformers model for efficient semantic sentence representation and clustering.

Transform and enhance conversational AI with Phi-4 for improved language model performance and safety.

Mistral-7B-Instruct-v0.3 is a fine-tuned model with extended vocabulary and function calling.

Mistral Small 3 achieves state-of-the-art performance among small language models with 24 billion parameters.

Mistral-Nemo-Instruct-2407 outperforms similar models by integrating multilingual data and improved architecture.

The Ministral-8B-Instruct-2410 model excels in multilingual tasks and code-related benchmarks, designed for on-device computing.

Qwen2.5 language models offer enhanced multilingual support, instruction-following, and long-context capabilities.

Qwen2.5-7B-Instruct-1M, a powerful language model, excels in long-context tasks with enhanced efficiency.

Qwen2.5-Coder excels in code generation, reasoning, and fixing with 128K context support.

Qwen2.5-Math improves multilingual math problem-solving with enhanced reasoning and computational accuracy.

QwQ-32B-Preview is an AI model with potential but requires improvements in reasoning.

Aya Expanse 8B is a multilingual model leveraging advanced research breakthroughs for optimized performance.

Aya 23 is a highly capable multilingual language model optimized for 23 languages and research use.

Aya model is a multilingual AI outperforming rivals, supporting 101 languages, enhancing communication globally.

Bilingual EXAONE 3.5 language models offer advanced, versatile text generation across device types.

EXAONE-3.0-7.8B-Instruct is an advanced bilingual AI model offering competitive benchmark performance.

Mistral Small 3 redefines small Large Language Models with 24B parameters, innovative capabilities, multilingual support.

This generative text model, Mistral-7B-v0.1, excels compared to similar models on multiple benchmarks.

Mistral-7B-Instruct-v0.2 offers refined instruction-based capabilities for effective text generation tasks.

The Mistral-7B-Instruct-v0.1 is a fine-tuned, instruction-following language model.

Mistral-Small-Instruct-2409 is a fine-tuned model designed for sequence tasks with 22B parameters.

Meta Llama 3 models excel at generating text through optimized transformer architecture for diverse applications.

Phi-3.5-mini offers a powerful multilingual AI model designed for effective text generation across various use cases with enhanced reasoning capabilities and long-context understanding.

Discover the Phi-3.5-vision model, a state-of-the-art, lightweight multimodal solution for image-text tasks.

Llama 2 offers scalable generative text models enhancing dialogue applications with large parameter sizes.

LLaVA is an advanced open-source chatbot designed for multimodal instruction research and language processing.

LLaVA is a fine-tuned open-source chatbot model using GPT for multimodal instruction-following.

Code Llama offers versatile models for code synthesis and understanding, designed for multiple programming needs.

TinyLlama efficiently pretrains a compact 1.1B Llama model, optimizing resources and enhancing adaptability.

StarCoder2-15B-Instruct is a self-aligned model optimizing code generation without human annotations.

DeepSeek-Coder-V2 excels in code tasks, supports 338 languages, and boasts advanced features.

Dolphin 3.0 R1, a flexible AI model, enables customized solutions for coding and general tasks.

Dolphin 3.0 Mistral 24B is an adaptable, user-controlled AI model for diverse applications.

Dolphin 3.0, a versatile AI model, empowers businesses with customizable and private solutions.

Dolphin 3.0 is an adaptable AI model focused on privacy, control, and customization.

Dolphin 3.0 is an advanced AI model offering personalized, general-purpose functionalities for diverse applications.

Introducing WizardLM-2, an advanced multilingual model with enhanced performance for complex tasks.

OLMo-2 models offer diverse language capabilities, perfect for cutting-edge state-of-the-art tasks.

OLMo 7B Instruct enhances language model science, excels at question answering, with open access.

OLMo is an open-source language model series by AI2, optimized for language model research.

OLMo 7B April 2024 improves language model performance using updated training techniques and datasets.

The OLMo 1B July 2024 model is enhanced with advanced dataset training, showing substantial performance improvements.

OLMo-2 models excel in diverse tasks using advanced training techniques and a versatile dataset.

Nous-Hermes-13b is a cutting-edge language model outperforming in long responses and task accuracy.

Nous-Hermes-Llama2-13b is optimized for long, accurate responses without censorship, outperforming predecessors.

Nous Hermes 2 - Yi-34B sets new standards with exceptional benchmark performance and usability.

Nous Hermes 2 on SOLAR 10.7B achieves enhanced performance on various AI benchmarks, closer to Yi-34B.

Nous Hermes 2 Mistral 7B DPO model exhibits improved performance across multiple benchmarking tests.

Nous-Hermes-Llama2-7b excels with long responses and reduced hallucination using diverse datasets.

C4AI Command R+ is a multilingual, advanced AI model trained for reasoning, summarization, and question answering using 104 billion parameters.

C4AI Command-R is a 35 billion parameter, multilingual generative model optimized for various tasks.

C4AI Command R 08-2024 is a multidimensional AI model excelling in multilingual tasks, reasoning, and citations.

Yi-1.5 enhances performance in language tasks, coding, and reasoning using extensive training.

Yi-Coder offers advanced open-source code models supporting 52 languages with exceptional performance and efficiency.

Yi-Coder delivers top coding performance using models under 10 billion parameters for 52 languages.

Explore Yi's next-gen, bilingual open-source models excelling in various benchmarks globally.

Yi's innovative bilingual language models excel in large-scale multilingual benchmarks, offering top-tier performance across tasks.

Yi series models are powerful bilingual, open-source AI language models, excelling in various NLP tasks.

Codestral-22B-v0.1 excels in code generation, utilizing 80+ programming languages like Python and Java.

Codestral Mamba 7B is a leading open code model, excelling in multiple benchmarks.

Sailor2 offers a multilingual language model for 15 South-East Asian languages, supporting diverse applications.

Sailor2 develops multilingual language models for Southeast Asia with accessible advanced technology.

Vicuna is a chat assistant fine-tuned from LLaMA, intended primarily for research purposes.

Vicuna is an advanced chat assistant model designed for research in NLP, AI, and ML.

Vicuna, an AI chat assistant, excels in research on language models and chatbots, benefiting NLP enthusiasts.

Vicuna is a chat assistant model, fine-tuned on user-shared conversations for NLP research.

Vicuna is a chat assistant trained on LLaMA for research in AI and NLP conversations.

Vicuna is a fine-tuned LLaMA model focused on chatbot research using user-shared conversations.

The Nemotron-Mini-4B-Instruct is an optimized language model for roleplay, Q&A, and function calls.

Falcon3-1B-Instruct is a versatile model excelling in reasoning, language, and inquiry tasks.

Falcon-7B-Instruct is a powerful, optimized model for chat and instruction tasks.

WizardCoder models significantly advance code language models, reaching state-of-the-art results in evaluations.

Explore WizardCoder, a leading Code Large Language Model with exceptional performance across multiple benchmarks.

Code Llama offers versatile models for code synthesis and understanding across various programming tasks.

DeepHermes 3, enhanced with advanced reasoning, offers flexible AI interaction and improved user alignment.

OpenThinker-7B excels in performance and open-source accessibility across multiple evaluation metrics.

Uncensored DeepSeek model version created via abliteration to enhance LLM functionalities.

DeepSeek-R1-Distill-Qwen-7B-abliterated-v2 offers an uncensored, refined language model version.

Velvet-2B, an Italian language model, excels in diverse linguistic applications, promoting ethical AI use.

SmolLM2 offers efficient, compact language models excelling in diverse tasks and instruction-following.

Finetuned Qwen2.5-Math model enhances mathematical problem-solving capabilities with extended context length.

QVikhr-2.5-1.5B-Instruct-r is a bilingual language model specialized in Russian math datasets.

ReaderLM-v2 efficiently converts HTML into markdown or JSON, supporting 29 languages with 512K token handling.

Llama 2 offers advanced, scalable language models with enhanced performance and ethical considerations.

Llama-3.1-Tulu-3 models excel in diverse tasks with advanced training techniques and open-source data.

OREAL mathematical reasoning models excel with innovative reinforcement learning, achieving significant accuracy improvements.

DeepSeek-llama3.1-Bllossom-8B enhances multilingual AI performance, notably improving Korean language inference.

DeepSeek-R1-Distill-Llama-8B-abliterated is an uncensored model offering advanced AI capabilities.

YuE is an innovative music generation model transforming lyrics into complete songs, supporting diverse genres and languages.

DeepSeek-R1-Distill-Qwen-32B-Japanese is a Japanese language model designed for advanced text generation.

Llama-Krikri-8B-Instruct enhances Greek text generation through extensive pretraining and bilingual capabilities.

Fino1-8B is a financial reasoning model fine-tuned for enhanced performance on specified tasks.