{"name":"DeepInfer","repo":"deepseek-ai/DeepInfer","url":"https://midorreal.com/project/deepseek-ai-deepinfer","repository":"https://github.com/deepseek-ai/DeepInfer","description":"DeepInfer: High-performance, large-scale inference framework for DeepSeek models","facts":{"license":{"value":"Apache-2.0","source":"https://github.com/deepseek-ai/DeepInfer","checked":"22 Nov 2026"},"language":{"value":"Python","source":"https://github.com/deepseek-ai/DeepInfer","checked":"22 Nov 2026"},"releases":{"value":"0","source":"https://github.com/deepseek-ai/DeepInfer/releases","checked":"19 Nov 2026"},"contributors":{"value":"5","source":"https://github.com/deepseek-ai/DeepInfer/graphs/contributors","checked":"19 Nov 2026"},"archived":{"value":"no","source":"https://github.com/deepseek-ai/DeepInfer","checked":"22 Nov 2026"}},"stars":26304.0,"writeup":{"what_it_is":"DeepInfer is an open-source, high-performance inference framework for DeepSeek models. It provides offline inference with features for serving at scale, including optimized kernels, parallelism options, and batch-invariant inference.","audience":"Developers serving DeepSeek models at large scale on NVIDIA SM100 or Huawei Ascend 950 hardware.","claims":[{"kind":"specific","claim":"Supports DeepSeek-V4.1-Flash as a model.","excerpt":"Supported models:\n\n- DeepSeek-V4.1-Flash","status":"not_checked"},{"kind":"specific","claim":"Runs on NVIDIA SM100 GPUs and Huawei Ascend 950 NPUs.","excerpt":"- NVIDIA SM100 GPUs\n- Huawei Ascend 950 NPUs","status":"not_checked"},{"kind":"specific","claim":"Provides batch-invariant inference that keeps results identical across different batch compositions.","excerpt":"keep model results identical across different batch compositions","status":"not_checked"},{"kind":"specific","claim":"Supports structured outputs constrained to a regex, JSON schema, or EBNF grammar.","excerpt":"constrain generation to a regex, JSON schema, or EBNF grammar","status":"not_checked"},{"kind":"specific","claim":"Reports measured benchmarks on 4 SM100 GPUs for prefill and 64 for decode.","excerpt":"Measured on 4 SM100 GPUs for prefill and 64 for decode","status":"not_checked"}],"alternatives":[],"written_by":"AI, from the project's README","date":"2026-10-21"},"why_now":null,"owner_supplied":null,"score":{"verdict":"overhyped","hype":86,"reality":51,"gap":36,"momentum":null,"confidence":85,"version":"score-2.0","rule":"verdict-2.0","calculated_at":"2026-11-22T07:00:00+00:00","verdict_source":"rule"},"early_read":null,"changes":[{"date":"2026-10-16T07:47:23+00:00","text":"Repository created","source":"https://github.com/deepseek-ai/DeepInfer"}],"cite":"DeepInfer is a GitHub project at https://github.com/deepseek-ai/DeepInfer written mainly in Python, described there as \"DeepInfer: High-performance, large-scale inference framework for DeepSeek models\".","attribution":"Data from Mid or Real, https://midorreal.com/project/deepseek-ai-deepinfer, checked 22 November 2026. Attribution with a link is required.","still_real":null}