Unverified50% confidenceFactExact time

llama.cpp was released by Georgi Gerganov in January 2023 and uses a self-developed GGML tensor library rather than PyTorch or JAX.

1

Sources

50%

Confidence

Long-term

Relevance

7/24/2026

First Seen

Sources

小米静默发布MiMo-V2.5-DFlash：300B模型推理速度或将翻倍

redditr/LocalLLaMA7/12/2026

Related Claims

Unverifiedggml 由保加利亚工程师 Georgi Gerganov 于 2022 年创建，是 llama.cpp、whisper.cpp 等本地推理项目背后的底层张量计算库80% similar Verifiedllama.cpp是由开发者Georgi Gerganov于2023年发起的开源项目，使用纯C/C++实现了LLaMA系列模型的推理79% similar VerifiedThe GGUF format was designed by Georgi Gerganov, the founder of the llama.cpp project.75% similar Partially VerifiedLlamaCPP was initiated by developer Georgi Gerganov in March 2023 as a personal project to run LLaMA models on a MacBook.67% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/608490

API

curl https://kongchang.com/api/v1/knowledge/claims/608490

MCP

get_claim(id=608490)