Вход на сайт

Просмотр новости

Найдите то, что Вас интересует

xAI Launches Grok 4.6 to Take On GPT-5: Here is What’s New

Дата публикации: 12-08-2026 18:51:50

SpaceXAI has just rolled out Grok 4.6, delivering major upgrades in long-running AI agents and visual project building at competitive pricing.
The post xAI Launches Grok 4.6 to Take On GPT-5: Here is What’s New first appeared on iPhone in Canada.

Основное содержимое страницы с новостью.

xAI has just announced the release of Grok 4.6 to take on the likes of OpenAI’s GPT-5, introducing significant upgrades aimed at handling complex, multi-step agentic tasks and interactive visual projects.

Comparison table of AI benchmark scores across Grok model versions (4.6 High highlighted). Rows include AA Intelligence Index, GDPVal-AA v2, CursorBench v3.2, DeepSWE v1.1, and more; bold numbers indicate best results per column across Grok 4.6, 4.5, GPT-5.6 Solo Max, and Fable 5 Max.

The new flagship model is available immediately through the company’s API, the Cursor code editor, and the proprietary Grok Build environment.

The release builds directly on the foundation established by Grok 4.5, maintaining the core 1.5 trillion parameter architecture while executing a much longer supplemental training pass. SpaceXAI utilized curated model-generated reasoning data, technical engineering datasets, and refined reinforcement learning techniques to sharpen the model’s problem-solving skills.

According to official performance documentation released by the company, Grok 4.6 matches OpenAI’s GPT-5.6 Sol on the Artificial Analysis Intelligence Index with a score of 61. On specialized tests, the model posted notable gains over its predecessor.

On the GDPVal-AA knowledge work benchmark, Grok 4.6 scored 1753 points, outperforming competing models. It also demonstrated impressive growth on coding evaluations, reaching 69.9 percent on CursorBench v3.2 and 65.9 percent on the DeepSWE v1.1 software engineering test, up from 54 percent in Grok 4.5.

A primary focus for Grok 4.6 is its capacity to sustain focus across extended workflows without losing context or derailment. When tasked with building web applications or software tools, the model can conduct background research in unfamiliar domains, set up project structures, implement core user interactions, and refine code based on user feedback.

The model also features an expanded 500,000-token context window, allowing developers to upload massive codebases or extensive technical documents into a single prompt session. For visual design tasks, Grok 4.6 establishes visual aesthetics and layout hierarchies in its initial response, making it easier for users to iterate on working software prototypes in real time.

Commenting on the launch, CEO Elon Musk called the release “Smart, fast & amazing bang for buck!”

Схожие новости

#Наименование новостиТональностьИнформативностьДата публикации
1xAI Debuts ‘Grok Bot’ AI Teammates You Can Give Real Work To06.6112-08-2026
2SpaceXAI unveils Grok 4.5 with faster coding, reasoning, and productivity features019.0209-07-2026
3SpaceXAI Will Release Grok 4.6 in 2 Weeks and Grok 4.7 in 4 Weeks015.324-07-2026
4SpaceXAI выпустила Grok 4.5 — конкурента Opus и GPT, но значительно дешевле5709-07-2026
5Grok 4.5: SpaceXAI выпустила недорогую ИИ-модель для программирования0709-07-2026
6SpaceXAI представила ИИ-модель Grok 4.5. Она быстрее и дешевле конкурентов0509-07-2026
7SpaceX lance Grok Bot, la nouvelle IA qui veut travailler à votre place013.8512-08-2026
8Artificial Analysis: Новая ИИ-модель Grok 4.6 превосходит конкурентов по ряду параметров019.9713-08-2026
9SpaceXAI представила модель Grok 4.6 с упором на автономные ИИ-агенты020.5113-08-2026
10SpaceXAI запускает Grok Bot: армию ИИ-агентов для рутины и работы в команде010.412-08-2026

Классификация: Пресс-релизы. Схожих патентов: 0. Схожих новостей: 10. Тональность: 0. Информативность: 10.58. Источник: www.iphoneincanada.ca.