Discover our Integrated AI Models

Excellent for coding, data analysis, and complex reasoning tasks with significant improvements in performance over earlier models.

Fast and efficient Claude model optimized for quick responses with strong reasoning capabilities.

Anthropic's most intelligent model for complex tasks with enhanced reasoning and coding capabilities.

Anthropic's most advanced model with industry-leading performance in coding, reasoning, and problem-solving with extended thinking capabilities.

Highest intelligence across most tasks with exceptional agent and coding capabilities with advanced thinking capabilities

This is the full version of DeepSeek R1, it is slower but more accurate.

Specialized reasoning model combining speed with advanced problem-solving abilities for complex tasks.

Updated version of DeepSeek V3.1 with improved language consistency and agent capabilities. Large hybrid reasoning model (671B parameters, 37B active) supporting both thinking and non-thinking modes with enhanced tool use, code generation, and reasoning efficiency.

DeepSeek V3.2 is the latest version with improved language consistency and enhanced capabilities.

Powerful multimodal model balancing performance and speed with native tool use and structured output capabilities.

Ultra-lightweight Gemini model for cost-effective, high-speed inference with massive context window.

Gemini 3 Flash is Google's next-generation fast model with improved performance and efficiency for rapid inference tasks.

Gemini 3 Pro Preview is Google's next-generation flagship model with enhanced reasoning, coding, and multimodal capabilities.

Google's most advanced open source model with 27B parameters, optimized for complex tasks.

Foundation model designed for intelligent agents with 355 billion total parameters and 32B active parameters, optimized for agentic tasks and complex reasoning.

Compact and efficient foundation model designed for intelligent agents with 106 billion total parameters and 12B active parameters. Optimized for agentic tasks, reasoning, and coding with excellent performance-to-cost ratio.

Advanced foundation model with expanded 200K context window, superior coding performance, advanced reasoning capabilities, and enhanced tool use. Improved performance in agentic tasks, front-end development, and refined writing with better human preference alignment.

Cost-effective multimodal model with excellent language capabilities, coding skills, and strong performance on academic benchmarks.

The smallest and fastest model in the GPT 5 series.

Powerful multimodal model with excellent capabilities in language understanding, visual reasoning, and complex tasks.

ChatGPT-4o is the version of GPT-4o that powers ChatGPT legacy model. Creative model for light work.

Open-weight 117B-parameter Mixture-of-Experts language model with 500 tokens/sec speed, designed for high-reasoning, agentic, and general-purpose production use cases with configurable reasoning depth.

Open-weight 21B parameter model with 1,000 tokens/sec speed and Mixture-of-Experts architecture, optimized for ultra-fast inference and agentic capabilities.

Multimodal model with strong performance on visual tasks and conversational capabilities.

Grok 3 Mini is a high-quality language model with exceptional performance for real-time applications and complex tasks.

Grok 4 is Grok's latest and most advanced model, it outperforms the best models in certain benchmarks.

Fast variant of Grok 4 optimized for reasoning with a 2M-token context window.

Fast non-reasoning variant of Grok 4.1 with a 2M-token context window and improved efficiency.

Fast, code-oriented model with large context window for software engineering tasks.

State-of-the-art mixture-of-experts model with 32B activated parameters and 1T total parameters, specifically designed for agentic intelligence, tool use, and autonomous problem-solving.

High-quality language model with exceptional throughput for real-time applications and complex tasks.

Llama 4 Scout is a high-quality language model with exceptional performance for real-time applications and complex tasks.

Advanced multimodal model from MiniMax with strong reasoning and tool use capabilities.

Advanced reasoning model with powerful capabilities for complex problem-solving and engineering tasks.

Qwen3's enhanced thinking model with scaled thinking capabilities, improving reasoning performance while maintaining high-quality instruction following and tool use.

Qwen3's most advanced agentic code model with exceptional performance on coding, agentic browser-use, and tool calling with 480B parameters and 35B active parameters.

QwQ is a large language model that is optimized for reasoning and has a 32B parameter count.

Specialized model for real-time web search with excellent information retrieval capabilities.

Enhanced web search model with improved accuracy and deeper analysis of online information.

Premium reasoning model with enhanced web search for solving complex problems with online information.

Advanced reasoning capabilities combined with comprehensive web search for complex research tasks.