{"name":"dynamo","repo":"ai-dynamo/dynamo","url":"https://midorreal.com/project/ai-dynamo-dynamo","repository":"https://github.com/ai-dynamo/dynamo","description":"A Datacenter Scale Distributed Inference Serving Framework","facts":{"language":{"value":"Rust","source":"https://github.com/ai-dynamo/dynamo","checked":"22 Nov 2026"},"homepage":{"value":"https://docs.nvidia.com/dynamo/latest","source":"https://docs.nvidia.com/dynamo/latest","checked":"22 Nov 2026"},"last_release":{"value":"v1.5.1-kimi-k3-post.1, 18 Nov 2026","source":"https://github.com/ai-dynamo/dynamo/releases","checked":"22 Nov 2026"},"releases":{"value":"73","source":"https://github.com/ai-dynamo/dynamo/releases","checked":"19 Nov 2026"},"contributors":{"value":"512","source":"https://github.com/ai-dynamo/dynamo/graphs/contributors","checked":"19 Nov 2026"},"archived":{"value":"no","source":"https://github.com/ai-dynamo/dynamo","checked":"22 Nov 2026"}},"stars":8545.0,"writeup":{"what_it_is":"Dynamo is an open-source orchestration layer that sits above inference engines such as SGLang, TensorRT-LLM, and vLLM. It coordinates them across multiple GPUs and nodes, with features including disaggregated serving, KV-aware routing, and automatic scaling.","audience":"Teams serving LLMs across multiple GPUs or nodes","claims":[{"kind":"specific","claim":"Separates prefill and decode into independently scalable GPU pools","excerpt":"Separates prefill and decode into independently scalable GPU pools","status":"not_checked"},{"kind":"specific","claim":"Routes requests based on worker load and KV cache overlap","excerpt":"Routes requests based on worker load and KV cache overlap","status":"not_checked"},{"kind":"specific","claim":"Reports 7x higher throughput per GPU for DeepSeek R1 on GB200 NVL72 compared with B200 without Dynamo","excerpt":"**7x** higher throughput per GPU","status":"not_checked"},{"kind":"specific","claim":"Reports 2x faster time to first token with KV-aware routing on Qwen3-Coder 480B","excerpt":"**2x** faster time to first token","status":"not_checked"},{"kind":"specific","claim":"Offers a zero-config deployment mode (DGDR) in beta where model, hardware, and SLA are specified in one YAML","excerpt":"Specify model, HW, and SLA in one YAML","status":"not_checked"}],"alternatives":[],"written_by":"AI, from the project's README","date":"2026-11-21"},"why_now":null,"owner_supplied":null,"score":{"verdict":"real","hype":67,"reality":92,"gap":-25,"momentum":1.3,"confidence":85,"version":"score-2.0","rule":"verdict-2.0","calculated_at":"2026-11-22T08:00:00+00:00","verdict_source":"rule"},"early_read":null,"changes":[{"date":"2026-11-18T02:29:37+00:00","text":"Latest release v1.5.1-kimi-k3-post.1","source":"https://github.com/ai-dynamo/dynamo/releases"},{"date":"2025-03-18T03:37:26+00:00","text":"First release","source":"https://github.com/ai-dynamo/dynamo/releases"},{"date":"2025-03-03T18:40:07+00:00","text":"Repository created","source":"https://github.com/ai-dynamo/dynamo"}],"cite":"dynamo is a GitHub project at https://github.com/ai-dynamo/dynamo written mainly in Rust, described there as \"A Datacenter Scale Distributed Inference Serving Framework\".","attribution":"Data from Mid or Real, https://midorreal.com/project/ai-dynamo-dynamo, checked 22 November 2026. Attribution with a link is required.","still_real":null}