Models46 articles
Kimi K2 Instruct
Kimi K2 Instruct is a 1-trillion-parameter Mixture-of-Experts (MoE) language model developed by Moonshot AI, specifically optimized for agentic intelligence and multi-step tool execution. It supports a 256,000-token context window and is positioned as a cost-effective, open-weights alternative to proprietary frontier models like GPT-4o.
Llama 2
Llama 2 is a family of pretrained and fine-tuned large language models released by Meta AI in July 2023, offering parameter sizes up to 70 billion. Developed as an open-weights alternative to proprietary models, it features a 4,096-token context window and specialized optimizations for dialogue and safety.
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.
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.
Llama 4 Maverick
Llama 4 Maverick is a specialized iteration of Meta's Llama 4 large language model series, optimized for autonomous agentic workflows, multi-step reasoning, and precise tool-use. It utilizes a sparse Mixture-of-Experts (MoE) architecture to balance high-level intelligence with inference efficiency.
Claude Opus 3
Claude 3 Opus is the flagship large language model of Anthropic's Claude 3 family, designed for high-level reasoning, complex analysis, and multimodal input processing. Released in March 2024, it was the first model to surpass GPT-4 on several key industry benchmarks before being succeeded by the Claude 3.5 and 4 series.
Claude Opus 4.0
Claude Opus 4.0 is Anthropic's flagship large language model featuring a hybrid reasoning architecture and an extended thinking mode for complex analytical tasks. It is optimized for autonomous software engineering and long-horizon research, and is the first model to implement AI Safety Level 3 protocols.
Claude Opus 4.6
Claude Opus 4.6 is Anthropic's frontier large language model designed for high-complexity cognitive tasks, featuring a sparse Mixture-of-Experts architecture and a 1-million-token context window. It is optimized for enterprise-level reasoning, autonomous agentic workflows, and sophisticated multimodal analysis of text, images, and video.
Claude Sonnet 3.5
Claude 3.5 Sonnet is a mid-tier large language model developed by Anthropic that balances high-level reasoning with operational speed. Released in June 2024 and upgraded in October 2024, it introduced advanced capabilities in coding, visual processing, and an experimental 'computer use' feature for interacting with standard desktop interfaces.
Claude Sonnet 3.7
Claude 3.7 Sonnet is a multimodal large language model developed by Anthropic, released in February 2025. It is the first "hybrid reasoning model" featuring an "extended thinking" mode designed for complex software engineering and mathematical 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.
Claude Sonnet 4.0
Claude Sonnet 4.0 is a frontier large language model developed by Anthropic that balances high-level intelligence with operational speed. It is optimized for software engineering and autonomous agentic workflows through a unique hybrid reasoning architecture.
Claude Sonnet 4.5
Claude Sonnet 4.5 is a mid-tier large language model in the Claude 4.5 family developed by Anthropic, optimized for complex software engineering and autonomous agentic workflows. It features a 200,000-token context window and advanced computer control capabilities.
Claude Sonnet 4.6
Claude Sonnet 4.6 is a mid-tier large language model in Anthropic's Claude 4.6 family, designed to balance high-speed processing with advanced reasoning and agentic capabilities. Released in February 2026, it features a 1-million-token context window and native multimodality, supporting complex workflows in software development and data analysis.
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.
Devstral Small 2505
Devstral Small 2505 is a 24-billion parameter large language model developed by Mistral AI and All Hands AI, specifically optimized for agentic software engineering and autonomous coding tasks. Released under the Apache 2.0 license, it features a 131,072-token context window and is designed to operate as a reasoning engine within agentic scaffolds like OpenHands.
Gemini 2.5 Flash Lite
Gemini 2.5 Flash Lite is a high-efficiency multimodal large language model developed by Google DeepMind, optimized for low-latency performance and cost-effective scaling across high-volume tasks. It utilizes a sparse Mixture-of-Experts architecture and supports a 1-million-token context window for processing text, audio, images, and video.
Gemini 3.1 Pro
Gemini 3.1 Pro is a multimodal large language model developed by Google DeepMind, designed for advanced reasoning, scientific knowledge, and autonomous agentic tasks. It features a unique "dynamic thinking" mechanism and a 1-million-token input context window with significantly expanded output limits.