Models

Kimi Kimi K2 Thinking Turbo model.

Kimi Kimi Latest model.

OpenAI Dall E 3 model.

OpenAI Text Embedding 3 Small model.

OpenAI Text Embedding 3 Large model.

OpenAI DALL-E 3 model.

OpenAI DALL-E 2 model.

Faster, more cost-efficient version of GPT-5 for well-defined tasks and precise prompts with 400K context window.

Enhanced reasoning model using additional compute for consistently better answers on hard math, science, and coding problems.

Google's latest image generation model (Gemini 3 Pro Image). Built for studio-quality precision with up to 4K resolution, allowing you to edit specific parts of an image through conversational turns.

Google's most advanced video generation model. Produces 4K cinematic video with native audio generation — sound effects and music are generated to match the visual action perfectly.

xAI's frontier reasoning model with 2M context window and step-by-step thinking for complex problems.

Non-reasoning variant of Grok 4.1 optimized for instant responses with 2M context and agentic tool calling.

Distilled version of DeepSeek R1 based on Qwen 2.5 32B architecture. Efficient reasoning at a fraction of the cost.

Kimi K2 with native thinking mode for step-by-step reasoning and 200+ sequential tool calls with 256K context.

Anthropic's latest flagship model with 1M token context, agent teams capability, and advanced agentic coding.

OpenAI's latest and most capable model with advanced reasoning, multimodal understanding, and improved instruction following.

Fast and efficient Gemini 3 model that outperforms 2.5 Pro while being 3x faster. Great balance of quality and speed with 1M token context.

xAI's dedicated image generation model (Aurora). Excels at photorealistic rendering and precisely following text instructions.

Moonshot AI's open-source multimodal model with 1T total parameters (32B active). Supports vision, agentic capabilities, and Agent Swarm coordination.

OpenAI's most advanced model with breakthrough reasoning capabilities, multimodal understanding, and improved instruction following.

Google's most capable model with state-of-the-art reasoning, 1M token context, and multimodal understanding across text, image, video, and audio.

Anthropic's latest model with excellent balance of intelligence and speed. Excellent for complex tasks requiring careful reasoning.

Fast and affordable Claude model with strong coding, reasoning, and tool use capabilities for high-throughput tasks.

Fast and efficient model balancing quality and speed with multimodal input support and thinking capabilities.

Moonshot AI's latest model with strong multilingual and reasoning capabilities routed through NVIDIA NIM.

High-performance model balancing intelligence and speed. Excellent for complex tasks requiring careful reasoning.

Alibaba's flagship model with 235B total parameters, strong multilingual support, and competitive reasoning.

Compact reasoning model optimized for speed and efficiency on complex tasks requiring chain-of-thought.

Advanced reasoning model with deep chain-of-thought capabilities for complex problem solving across math, coding, and science.

Powerful thinking model with built-in reasoning capabilities. Excels at math, coding, and complex multi-step problems.

Cohere's latest model optimized for enterprise RAG, tool use, and complex business workflows with strong grounding.

Qwen's reasoning model with 32B parameters optimized for complex problem solving with chain-of-thought.

Efficient and affordable model for lightweight tasks with strong multilingual support and fast response times.

Anthropic's most powerful model excelling at complex analysis, nuanced writing, advanced mathematics, and coding.

Strong general-purpose model with real-time information access and a conversational style.

Compact and efficient model for everyday tasks with fast response times and real-time knowledge.

Fast and efficient model for everyday tasks with multimodal input support and competitive quality at low cost.

Advanced reasoning model with transparent chain-of-thought. Competes with top models on math and coding at a fraction of the cost.

Specialized code generation model supporting 80+ programming languages. Optimized for code completion, generation, and explanation.

Versatile general-purpose model great for everyday questions, writing, and analysis with extremely competitive pricing.

Mistral's flagship model with strong multilingual support, function calling, and JSON mode for structured outputs.

Fastest Claude model optimized for near-instant responses. Ideal for chat, classification, and high-throughput tasks.

Small, fast, and affordable model for lightweight tasks. Great for classification, extraction, and simple generation.

Enterprise-grade model optimized for retrieval-augmented generation (RAG), tool use, and complex business workflows.