Developed with SGLang, Atlas Inference surpasses leading AI companies in throughput and cost, running DeepSeek V3 & R1 faster than DeepSeek themselves.

Despite its smaller size, DeepSeek-R1-0528-Qwen3-8B beats Google’s Gemini 2.5 Flash on a tough math test called AIME 2025 and ...

Developed with SGLang, Atlas Inference surpasses leading AI companies in throughput and cost, running DeepSeek V3 & R1 faster than DeepSeek ... dramatically reduces GPU and server requirements ...