xAI has announced Grok 4.1 Fast, its most powerful tool-calling model to date, now available via the xAI API alongside a new Agent Tools API that grants agents access to real-time X data, web search, code execution, and more. Grok 4.1 Fast features a truly massive 2 million-token context window, enabling it to reason and complete complex, agentic tasks such as customer support and financial workflows with high accuracy and speed. Trained through reinforcement learning in simulated environments across diverse domains, it achieves a perfect score on the τ²-bench Telecom benchmark, backed by independent evaluation. The model also delivers state-of-the-art tool-calling capabilities, posting 72 % overall accuracy on the Berkeley Function Calling v4 benchmark, while reducing hallucination rates by half compared to Grok 4 Fast.
Also Read: Rockwell Automation Adds Edge GenAI with NVIDIA Nemotron
The Agent Tools API allows Grok to operate as a fully autonomous agent, orchestrating browser-based research, social media sentiment tracking, code execution, and document retrieval all on xAI’s infrastructure. Two variants are offered grok-4-1-fast-reasoning (for maximal intelligence) and grok-4-1-fast-non-reasoning (for instant responses) with a special free access window: both models and agent tools are free via OpenRouter until December 3, 2025. Pricing after that is competitive: input tokens cost $0.20 per million (or $0.05 for cached), output tokens $0.50 per million, and tool calls from $5 per 1,000 successful invocations. This launch positions Grok 4.1 Fast as a serious contender for enterprise-grade intelligent agents, making it easier for developers to build production-ready systems without the complexity of managing API keys, rate limits, or retrieval pipelines.
