Open-Source Model Reaches 2.8 Trillion Parameters, Claims World's Largest
Decision Brief
What changedAn open-source model claims to have 2.8 trillion parameters, making it the largest and strongest open-source model globally.
Why it matters2.8 trillion parameters is a new record for open-source models, ushering in a trillion-parameter era, but actual performance and resource demands remain concerns.
Who should careOpen-source model users
Affected stackLlama
Source confidenceMedium · Reliable media or first-hand reporting
According to Dongfang Caifu, an open-source model announced it has reached 2.8 trillion parameters, claiming to be the world's largest and most powerful open-source model. This marks the first time an open-source model surpasses 2 trillion parameters, far exceeding previous models like Llama 3.1 405B. For developers and researchers using open-source large models, this means potentially stronger capabilities but also significantly higher hardware and inference costs. A 2.8 trillion parameter model demands immense GPU memory and bandwidth, currently affordable only by very few organizations.
Summary basis: official / RSS sourceCompiled from the source scope noted above; the original remains authoritative.
Sources
- Google News:中国模型(中文)
Full-web discovery via Google News: Chinese-language coverage of 通义/DeepSeek/Kimi/智谱/豆包/MiniMax model updates.
- Google News:中国模型(中文)
留言
登入后即可留言,和其他 builder 交换实测心得。
还没有留言,抢头香。