Real-time multi-agent AI that debates itself to find truth.
Stop chatting with a single model; start consulting a council. Grok 4.2 introduces a native multi-agent architecture where four experts: Grok (Coordinator), Harper (Research), Benjamin (Logic/Code), and Lucas (Creative) work in parallel. They cross-check facts and debate conclusions in real-time before you see the answer. Built for "rapid learning," Grok 4.2 iterates weekly based on your feedback, slashing error rates to just 4.2% while staying an order of magnitude faster.
Batch API reduces token costs by 20-50% compared to real-time pricing, with most requests completing within 24 hours. Ideal for non-urgent workloads like data processing, content generation, and analysis where latency tolerance exists.
Grok 4.3 supports strong agentic tool calling with minimal hallucinations. Server-side tools include web search, X search, code execution, file attachments, and collections search (RAG). Developers can build autonomous agents that decide tool usage based on query complexity.
Imagine API supports image generation (1K-2K resolution), image editing, video generation (480p-720p), and video editing. Pricing starts at $0.002-0.07 per image and $0.01-0.07 per second of video, with industry-leading speed.
Grok 4.3 provides a unified API for text generation, image/video generation via Imagine API, and voice capabilities via Voice Agent API. Developers can build end-to-end AI applications with a single platform, reducing integration complexity and supporting multi-turn conversations, function calling, and structured outputs.
Voice Agent API enables real-time speech-to-speech conversations at $3.00/hour, with text-to-speech at $15.00/1M characters and speech-to-text at $0.10-0.20/hour. Supports custom voice cloning and 25-language transcription.