Orbit/Models/Grok 4

Grok 4

by xaireasoningflagship
Context
256K tokens
Input
$3/M
Output
$15/M
License
Proprietary
Released
Jul 2025

Overview

Grok 4 is xAI's fourth-generation language model, released on July 9, 2025. It features a 256,000-token context window, enabling the processing of extensive text inputs. The model supports multimodal inputs, including both text and images, and offers structured outputs. Notably, Grok 4 integrates real-time access to the X (formerly Twitter) platform, facilitating current-events Q&A capabilities. It also supports parallel tool calling and configurable reasoning efforts, allowing users to adjust the model's reasoning capabilities. Grok 4 is available through multiple providers, including Microsoft Azure AI, OpenRouter, Oracle Cloud Infrastructure, Replicate, Vercel AI Gateway, and xAI's own platform. The model is priced at $3.00 per million input tokens and $15.00 per million output tokens. As of May 15, 2026, Grok 4 has been deprecated, with a knowledge cutoff date of July 31, 2025. The model is licensed under xAI's proprietary terms. In benchmark evaluations, Grok 4 achieved an Intelligence Index score of 41.5, a Coding Index score of 40.5, and a Math Index score of 92.7. It scored 0.9 in the MMLU-Pro benchmark and 0.9 in the GPQA benchmark. The model also achieved a score of 0.2 in the Humanity's Last Exam (HLE) benchmark, 0.8 in the LiveCodeBench, and 0.9 in the AIME benchmark. These results position Grok 4 as a competitive model in the AI landscape, particularly for applications requiring extensive context processing and current-events awareness.

Strengths

reasoningcurrent-events awarenesslong context

Benchmarks

Math Indexmath92.7
Intelligence Indexreasoning41.5
Coding Indexcoding40.5
MMLU-Proknowledge0.9
GPQAreasoning0.9
LiveCodeBenchcoding0.8
HLEreasoning0.2