deepseek-ai/DeepInfer
View on GitHub ↗ ★ 25,730 Python
DeepInfer: High-performance, large-scale inference framework for DeepSeek models
6 mentions on Hacker News by 6 people: 1 story, 5 comments · first 2026-10-20 · latest 2026-11-03
20082026
-
Deep seek have released their inference engine. https://github.com/deepseek-ai/DeepInfer
comment by vorticalbox · 2026-11-03 · in OpenAI seems to be silently re-enabling model training on Codex chats -
…inference engine, revealing that they're still running on only H100s and Huawei Ascend chips ( https://github.com/deepseek-ai/DeepInfer/blob/main/README.md ). Anthropic will be running their 2027 Opus and Fable models on Vera Rubin, a chip…
comment by JacobAsmuth · 2026-11-03 · in Anthropic reveals $900B AI spending commitments ahead of IPO -
…155 trillion tokens per day, with overall revenue of US$38 per GPU-hour (at valley pricing)." https://github.com/deepseek-ai/DeepInfer/blob/main/doc/infer... So this implies they actually hit about double that during peak pricing.
comment by riknos314 · 2026-10-26 · in OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why? -
…just disabuse us of that notion? > …with overall revenue of US$38 per GPU-hour (at valley pricing) https://github.com/deepseek-ai/DeepInfer/blob/main/doc/infer...
comment by phoghed · 2026-10-23 · in Ask HN: If the models are so good, why doesn't OpenAI just start 100 companies? -
They use 3FS [0] also open sourced [1] [0] https://github.com/deepseek-ai/DeepInfer/blob/main/doc/infer... [1] https://github.com/deepseek-ai/3fs
comment by fotta · 2026-10-20 · in DeepInfer: High-performance inference framework for DeepSeek models