On December 17, according to Business Insider, in order to create an "artificial intelligence priority" office environment, it has been reported to the government that it has been able to do soMeta The use of artificial intelligence tools by competitors like Google and OpenAI is being expanded。

Several internal documents and posts obtained by the media show that the social media giant has been encouraging staff to integrate artificial intelligence tools into almost all work processes。
Meta's Chief Information Officer, Atitish Bannaki, stated to employees in a memorandum in June that one of the first tasks of the company was “to make artificial intelligence the core of our working methods”. The memorandum also describes a plan to combine Meta ' s self-developed series of models named after Llama with products from other enterprises。
In November, a Meta engineer disclosed in an internal post that the staff had been able to use the Google Gemini 3 Pro and Openai ChatGPT-5. The post also lists the artificial intelligence tools available to employees and their application scenarios, organized as follows:
| Tool Name | Application Scenario |
|---|---|
| Agentic Auto | A top-level smart body with tools and sub-intelligence |
| Advanced Auto | An experimental intelligence capable of handling more complex tasks |
| Integrated GPT-5 Thinking | A top-of-the-board smart body to produce better answers by extending the time of thinking |
| iLlama | A "fast" Llama 3.1 model to optimize the internal problems of Meta |
| Llama 4 Maverick | Llama 4 Maverick model without Meta internal data |
| GPT-4.1 | Advanced Language Model without Meta Internal Information |
| Gemini 3 Pro | Advanced Language Model without Meta Internal Information |
| Classic Agent | Replaced by Agenic Auto, coming down |
| Devmate | Artificial Intelligence Programming Aid for Integrated Development Environments |
Meta spokesmen confirmed this upgraded set of artificial intelligence tools and quoted earlier statements on the use of artificial intelligence technology: “As we all know, artificial intelligence is the focus of corporate development, and we are working to help employees do their daily work efficiently through artificial intelligence.”
In June this year, the social media giant eased restrictions to allow employees to use artificial intelligence models developed by competitors. It includes an internal coding tool called Devmate, which uses the Anthropic Claude model。
In a memorandum of June, Bannaki stated that Gemini and NotebookLM Pro of Google had also been extended throughout the company to help staff “to work more efficiently to create greater value”。
Meta has invested tens of billions of dollars in self-studying an artificial intelligence model for consumers, and employees can use an in-house artificial intelligence assistant called Metamate, built on its Llama model。
According to a pre-line internal announcement of the tool, Meta worked with the artificial intelligence start-up company Midjourney this summer to integrate its image generation technology into its own products and models; in October of the same year, the tool was officially opened to staff for “concept design and production” to accelerate the development of design processes and creative prototypes。
1AI notes that Gemini is not the only Google tool introduced by Meta. This summer, Meta fully migrated the in-house suite to Google Workspace, covering chat tools, e-mail, documents and cloud hard drives. In its memorandum of June, the company stated that the aim was “to unlock artificially intelligently driven functions” and to better adapt to its growing tool ecology。
In the field of engineering technology, Meta further opens up access to smart programming systems by adding Gemini 3 Pro to Google, while exploring new integration with the Openai Code Generation Command Line Tool (Codex CLI) and the Gemini Command Line Tool (Gemini CLI) to Google. A November memo shows that Meta's “realistic lab” executive, Mahe Saba, said to his staff: “Our strategic core is not tied to a particular solution, but rather focused on the final results - improving the efficiency of work, accelerating the progress of research and development, and providing staff with the best intellectual programming experience.”
Meta has also introduced interesting incentives to encourage employees to use and explore artificial intelligence technologies. At the beginning of this year, the company launched a small in-house game called "Level Up " , where employees are rewarded for having the flexibility to use artificial intelligence at work. Corporate management also linked staff performance to the effectiveness of artificial intelligence applications, rewarded staff members who could prove “value creation through artificial intelligence technology” during the year and planned to incorporate artificial intelligence application performance into the performance appraisal system by 2026。