Models46 articles
Qwen 3 30B A3B
Qwen 3 30B A3B is a 30-billion parameter large language model developed by Alibaba Cloud, utilizing a sparse Mixture-of-Experts (MoE) architecture that activates 3 billion parameters per token for high efficiency. It is designed to provide mid-sized model reasoning capabilities with the speed and cost profile of a much smaller system, supporting a context window of up to 262,144 tokens.
R1-0528 Turbo
R1-0528 Turbo is a high-efficiency large language model developed by DeepSeek AI, optimized for throughput and complex reasoning using a Mixture-of-Experts (MoE) architecture. It is designed to provide advanced logic and coding capabilities with significantly reduced computational overhead and API costs.
R1 Distill Llama 70B
DeepSeek-R1-Distill-Llama-70B is a 70.6-billion parameter open-weights reasoning model developed by distilling the capabilities of the larger DeepSeek-R1 into a Llama-3.3-70B foundation. It is specifically optimized for high-level mathematical reasoning, coding, and logical tasks using an explicit chain-of-thought processing mechanism.
Llama 3.3 70B Instruct Turbo
Llama 3.3 70B Instruct is a large language model developed by Meta AI that utilizes knowledge distillation from the Llama 3.1 405B model to provide flagship-level reasoning capabilities in an efficient 70-billion parameter architecture. It features a 128,000-token context window and is optimized for complex tasks such as agentic workflows, multilingual communication, and software development.
V3.2 Exp
V3.2 Exp is an experimental large language model developed by DeepSeek that introduced the DeepSeek Sparse Attention (DSA) mechanism to optimize long-context processing. Released in September 2025, it utilizes a Mixture-of-Experts architecture and serves as a hybrid model for both general-purpose instruction and complex reasoning.
V3-0324
DeepSeek V3-0324 is an open-weights Mixture-of-Experts (MoE) language model with 671 billion total parameters, optimized for high-level reasoning, coding, and technical tasks. Released in March 2025, it features innovations like Multi-head Latent Attention and Multi-Token Prediction to balance high performance with inference efficiency.
GPT-4o
GPT-4o is a multimodal large language model developed by OpenAI that natively processes and generates text, audio, and visual data within a single integrated neural network. It features significantly reduced latency compared to previous iterations, enabling real-time human-computer interactions such as live translation and interactive tutoring.
GLM-4.5
GLM-4.5 is a multimodal large language model developed by Zhipu AI, featuring a Mixture-of-Experts (MoE) architecture with 355 billion parameters. Released in August 2024, it is designed for high-performance bilingual tasks in Chinese and English with integrated vision and reasoning capabilities.
Grok 2 Vision
Grok 2 Vision is a multimodal large language model developed by xAI that integrates native visual processing with reasoning capabilities. Launched in August 2024, it enables users to analyze images, documents, and complex charts through the X platform and a dedicated API.
R1-1776
R1-1776 is a reasoning-focused large language model released by Perplexity in 2025, designed as a modified, uncensored version of DeepSeek-R1. It employs targeted post-training to eliminate geopolitical restrictions while maintaining high performance in logical reasoning and mathematics.
Grok 4 Fast
Grok 4 Fast is a multimodal large language model developed by xAI, optimized for high inference speed and cost-efficiency. Released in September 2025, it features a 2-million-token context window and is designed for high-throughput applications such as real-time search and document analysis.
GPT-4o Search Preview
GPT-4o Search Preview is an integrated real-time search functionality within OpenAI's multimodal GPT-4o model, designed to provide up-to-date information with direct source attribution.
Claude Haiku 4.5
Claude 4.5 Haiku is a high-speed, intelligence-dense large language model developed by Anthropic, optimized for high-volume automated tasks and real-time user interactions with a 200,000-token context window.
Qwen 3 Next 80B Instruct
Qwen 3 Next 80B Instruct is a high-efficiency large language model featuring an ultra-sparse Mixture-of-Experts (MoE) architecture and hybrid attention. It activates only 3 billion of its 80 billion parameters per inference step, enabling significant throughput advantages for long-context tasks up to 256,000 tokens.
Qwen 3 Next 80B Thinking
Qwen 3 Next 80B Thinking is a specialized reasoning large language model developed by Alibaba Cloud, featuring a high-sparsity Mixture of Experts architecture and an internal chain-of-thought mechanism. It is designed for complex tasks in mathematics, logic, and software engineering, operating with 80 billion total parameters but only 3 billion active per token.
Mistral Small 3.2 24B Instruct
Mistral Small 3.2 24B Instruct is an open-weight multimodal large language model featuring 23.6 billion parameters and a 128,000-token context window. Released by Mistral AI in June 2025, it is optimized for high-speed instruction following, coding proficiency, and visual document analysis.
Qwen 3 14B
Qwen 3 14B is a dense 14.8 billion parameter large language model developed by Alibaba Cloud, featuring a hybrid reasoning engine that allows toggling between thinking and non-thinking modes. Released in April 2025 under an Apache 2.0 license, it achieves performance parity with much larger previous-generation models in STEM, coding, and logical reasoning tasks.
Mistral Small 3.1 24B Instruct
Mistral Small 3.1 24B Instruct is a 24-billion parameter multimodal model by Mistral AI that balances computational efficiency with advanced reasoning and vision capabilities. It features a 128,000-token context window and is optimized for local deployment on consumer-grade hardware and enterprise-scale agentic workflows.