XAI launches Grok-Code-Fast-1: speed and power for developers

By Canuto

XAI launches Grok-Code-FAST-1, an artificial intelligence model optimized for agile, fast and economical programming, available for free for limited time on platforms such as Github Copilot and cursor.
***

`Grok-Code-Fast-1` stands out for its speed, with 190 tokens per second and cache optimizations that reach 90% hits.
It is versatile in languages such as TypeScript, Python, Java, Rust, C ++ and Go, ideal for daily development tasks.
It offers a competitive price: USD $ 0.20 per million input tokens and USD $ 1.50 per million output tokens.

A new standard in programming assisted by AI

Artificial intelligence (AI) has transformed programming, and XAI, a company focused on accelerating human scientific discovery, has just taken a bold step. On August 28, 2025, XAI presented Grok-Code-FAST-1, a reasoning model specifically for agent coding workflows. This model promises to be an agile and economical tool for developers, standing out for their speed and ability to handle complex tasks with minimal supervision.

Unlike other AI models, which can feel slow in iterative workflows, Grok-Code-Fast-1 was created from scratch with a new architecture. XAI compiled a pre-entry corpus rich in programming content and cured post-training data sets based on requests for extraction and real-world coding tasks. This approach guarantees that the model is relevant to the daily challenges of programmers.

The model is now available for free for limited time on associated platforms such as Github Copilot, Cursor, Cline, Roo Code, Kilo Code, OpenCode and Windsurf. This strategy seeks to democratize access to advanced AI tools, allowing developers of all levels to experiment with Grok-Code-FAST-1.

The collaboration with launch partners was key. XAI worked closely with these platforms to refine the behavior of the model, ensuring that it integrates perfectly with common tools such as GREP, terminals and file editors. This makes it ideal for integrated development environments (IDE).

Unprecedented speed

The speed is the distinctive seal of Grok-Code-Fast-1. With a capacity of 190 tokens per second (TPS), it exceeds competitors such as Gemini 2.5 Pro, GPT-5 and Claude Sonnet 4, according to measurements made through the APIs of each supplier. This performance is due to innovative techniques developed by XAI inference and supercomputing teams.

In addition, XAI optimized storage in Prompts cache, achieving success rates greater than 90% with its launch partners. This means that the model can respond quickly, even in workflows that require multiple calls to tools. For example, you can execute dozens of commands before a developer finishes reading the first paragraph of reasoning monitoring.

A practical example of this speed comes from a developer who created a prototype of battles simulator in less than a day using Grok-Code-Fast-1 in cursor. The key was to divide tasks into small and focused fragments, allowing rapid iterations. This approach maximizes the efficiency of the model and facilitates the creation of complex projects.

The community has praised this speed. Mario Rodriguez, product director of GitHub Copilot, stressed that “in initial tests, Grok Code Fast has demonstrated its speed and quality in agent coding tasks.” This feedback reinforces the potential of the model to transform the workflows of developers.

Versatility and economy

Grok-Code-Fast-1 is not only fast, but also versatile. Domates languages such as TypeScript, Python, Java, Rust, C ++ and Go, and can perform tasks that range from creating projects from scratch to correcting specific errors. His ability to answer questions about code bases and make surgical corrections makes it a valuable tool for programmers of all levels.

The model obtained a 70.8% score in the complete Swe-Bench-Verified subset, an evaluation standard for coding models. However, XAI acknowledges that benchmarks do not completely capture the user experience. Therefore, they combined public evidence with human evaluations, where experienced developers described the performance of the model in daily tasks.

In terms of cost, Grok-Code-Fast-1 is highly competitive. With a price of USD $ 0.20 per million input tokens, USD $ 1.50 per million starting tokens and USD $ 0.02 per million tokens in cache, it is an accessible option for individual equipment and developers. This price structure seeks to balance performance and affordability.

XAI also offers an Prompts engineering guide to maximize model performance. This guide includes practical tips to structure requests and optimize workflows, ensuring that users obtain the best possible results.

The future of the Grok Code family

Initially released under the name Sonic, Grok-Code-FAST-1 has been subject to multiple updates based on community feedback. XAI monitored communication channels and deployed new control points of the model to address initial comments. This iterative approach is central to Xai’s philosophy.

The company plans to launch frequent updates, with improvements that will arrive in days instead of weeks. A new variant of the model, which will support multimodal tickets, called parallel to tools and an extended context length, is already in training. These innovations promise to carry the coding assisted by AI to new levels.

XAI also invites the developer community to share honest feedback. This open dialogue is crucial to refine the model and ensure that it meets the needs of the real world. Developers can access the model through the XAI API, with the prices mentioned above.

The impact of Grok-Code-Fast-1 is already felt on the community. Its combination of speed, versatility and accessibility positions it as a transforming tool for modern programming. As Xai continues it to, the future of agentive coding seems brighter than ever.

Original image of Diariobitcoin, created with artificial intelligence, for free use, licensed under public domain.
This article was written by an AI content editor and reviewed by a human editor to guarantee quality and precision.

WARNING: Diariobitcoin offers informative and educational content on various topics, including cryptocurrencies, AI, technology and regulations. We do not provide financial advice. Cryptactive investments are high risk and may not be adequate for all. Investigate, consult an expert and verify the applicable legislation before investing. I could lose all its capital.

XAI launches Grok-Code-Fast-1: speed and power for developers