<p>➀ DeepSeek仅花费600万美元用于计算能力以训练其R1模型,展示了使用更少或较不先进的芯片也能获得高质量的人工智能结果。</p><p>➁ 根据聊天机器人提供商DocsBot的数据,使用DeepSeek-R1上传和下载1百万token的成本分别为55美分和2.19美元。</p><p>➂ 深度学习模型R1在多个行业基准测试中与竞争对手的表现相当,甚至略胜一筹。相比之下,其他美国公司如OpenAI和xAI的模型虽然具有更大的输出窗口,但在性能测试中表现较差且成本更高。</p>
Related Articles
- AI bubble conundrum29 days ago
 - Advanced Motion Sensing For Smart Glassesabout 1 month ago
 - AI Accelerator For Edge Devicesabout 2 months ago
 - Light Powered Chip Enhances AI Efficiencyabout 2 months ago
 - Generative AI That Works Faster On Small Devicesabout 2 months ago
 - Qualcomm adds UHF RFID to mobile processor2 months ago
 - Ed Goes For Humanoid Robots3 months ago
 - Building Future-Ready AI Hardware With Neuromorphic Computing And Sensing3 months ago
 - China forms AI alliances to cut U.S. tech reliance — Huawei among companies seeking to create unified tech stack with domestic-powered standardization3 months ago
 - Transformer Accelerator Brings Large AI Models To Devices3 months ago