Translated data: Meta象 has announced the open-source release of its XVERSE-7B general-purpose large model, including both the base version and the conversational edition. This model is fully open-source, free for commercial use, and is more compact compared to its predecessor, the XVERSE-13B. The 7B version can run on a single consumer-grade graphics card, with a minimum of just 6GB of VRAM required after inference quantization, significantly lowering the development threshold and inference costs. The model has performed exceptionally well in multiple authoritative tests and possesses strong capabilities in cognition, planning, reasoning, and memory, supporting over 40 languages.