Longer-running task performance
A larger base model and a longer reinforcement-learning run focused on difficult tasks that may take many hours to complete.
Grok 4.7
Grok 4.7 is an xAI model for coding and knowledge work, available via Grok Build, Cursor, the Grok API and more, from $2 per million input tokens.
Grok 4.7 for coding and knowledge work
Grok 4.7 is an AI model from xAI for software development and knowledge work. It is positioned for tasks that require sustained reasoning, including longer coding jobs, document and presentation creation, and professional workflows such as office, legal, clinical, electrical-engineering, and cybersecurity work.
Compared with Grok 4.6, the model uses a larger base model, longer reinforcement-learning training on harder tasks, and improved handling of extended context and self-checking. It is available through several developer and coding environments, including Cursor, Grok Build, the Grok API, third-party coding harnesses, model routers, and cloud platforms. The source lists token-based pricing starting at $2 per million input tokens and $6 per million output tokens, plus a faster variant at twice the output speed and twice the price.
Key features
Self-verification
Designed to check its own work more carefully, supporting workflows where intermediate results need review before completion.
Long-context handling
Improved management of longer context, which is relevant to extended coding sessions and multi-step knowledge-work tasks.
Grok Bot harness support
Native understanding of the Grok Bot harness for conversational tasks and general knowledge work.
Broad professional-task coverage
Benchmarked across software engineering, electrical engineering, office work, legal work, clinical reasoning, and cybersecurity-related tasks, with results differing by benchmark.
Safety and cybersecurity safeguards
A new safeguard stack focused on refusal and jailbreak resistance. The source reports 62.4% on LatchBio’s biosafety benchmark and 3.3% of risky dual-use prompts allowed through on HackerBench v0.3, while emphasizing access to legitimate security work.
Use cases
Long-running software development
Use it for software-engineering tasks that run for extended periods, where longer context and more deliberate work checking can help maintain continuity across steps.
Documents and presentations
Apply it to multi-hour office work such as preparing professional documents and presentations. The source reports improvements over Grok 4.6 on GDPval and AA Briefcase benchmarks.
Professional analysis
Use it as a general-purpose model for professional analysis across legal, clinical, and electrical-engineering tasks. Published benchmarks cover these areas, but results differ by task.
Cybersecurity defense
Use it for defensive cybersecurity research and legitimate security work. The model is designed to provide useful cyber capabilities while refusing risky or malicious dual-use requests.
Conversational and developer workflows
Access it through Grok Build, Cursor, or an API-connected coding workflow depending on whether the task is exploratory, conversational, or integrated into developer tooling.
Pros and Cons
Pros
- Designed for difficult, multi-hour coding and knowledge-work tasks.
- Includes improved self-verification and longer-context management compared with Grok 4.6.
- Available through Cursor, Grok Build, the Grok API, third-party coding harnesses, model routers, and cloud platforms.
- Publishes separate safety claims for benign cybersecurity utility, jailbreak resistance, and risky dual-use refusal behavior.
- Supports a free trial path through Grok Build and token-based API pricing.
Cons
- The available source does not document specific context-window limits, file formats, output schemas, collaboration controls, or deployment requirements.
- Benchmark results vary substantially by task, so the published scores do not establish uniformly leading performance across every workflow.
- The faster variant costs twice as much as the standard pricing level.
FAQ
Where can I use Grok 4.7?
Grok 4.7 is available in Cursor and Grok Build, and it can also be accessed through the Grok API, third-party coding harnesses, model routers, and cloud platforms. The source page also provides links for creating an API key and reading the xAI documentation.
Can I try Grok 4.7 for free?
The model is available to try for free in Grok Build. The source does not specify the limits or conditions of that access.
How is Grok 4.7 priced?
The listed starting prices are $2 per million input tokens and $6 per million output tokens. A fast variant provides twice the output speed at twice the price; the source does not provide further plan or usage limits.
What kinds of work is Grok 4.7 designed for?
Grok 4.7 is intended for coding and knowledge work, including longer-running software tasks, document and presentation creation, office work, electrical engineering, legal work, clinical reasoning, and cybersecurity work. Benchmark results indicate that performance varies by task, so the model should be evaluated against the requirements of a specific workflow.
What workflow details and outputs are documented?
The model can work through the Grok Bot harness and is described as better at verifying its own work and managing longer context. The source does not specify particular file formats, output schemas, collaboration controls, or deployment requirements.
Quick Facts
- Category
- AI model / developer tool
- Primary uses
- Coding and knowledge work
- Availability
- Cursor, Grok Build, Grok API, third-party coding harnesses, model routers, and cloud platforms
- Starting input price
- $2 per million tokens
- Starting output price
- $6 per million tokens
- Fast variant
- Twice the output speed at twice the price
Grok 4.7 Alternatives
ByteAsk
ByteAsk is a terminal-first AI coding agent for C and C++ that edits repos and verifies changes with compilers, debuggers, sanitizers, and tests.
Ghost
Ghost is a terminal-based AI assistant for chatting, code generation, and CLI tasks. Includes free models, supports Linux, macOS, Windows, and is open source.
Vi3ecode
Vi3ecode is a maintained development environment for AI coding agents, keeping project context, terminal, Git, memory, flows and review connected.
Lucid Train
Lucid Train is a local-first AI coding harness that turns repo code into architecture diagrams for coding agents. Desktop app and Rust CLI, local models or API keys.
MeetStream
MeetStream is a meeting bot API for Zoom, Google Meet, Microsoft Teams, and Webex. Record, stream, and analyze meetings with usage-based pricing and $5 free credit.
CreateOS Sandbox
CreateOS Sandbox is an isolated compute environment for running code and agent workloads in Firecracker micro-VMs with private networking and SDK, CLI, or MCP control.