{"type":"post","id":"hibana-cornell-led-inference-engine-for-local-agentic-55553a","slug":"hibana-cornell-led-inference-engine-for-local-agentic-55553a","title":"Hibana: Cornell-led inference engine for local agentic workloads on unified-memory devices like DGX Spark","description":"A research team introduced Hibana, an inference engine designed for agent workloads on unified-memory devices.","retrieval_nugget":"A research team introduced Hibana, an inference engine designed for agent workloads on unified-memory devices. Its arXiv abstract reports up to 3.4 times more agentic tasks per hour and 3.9 times higher decoding throughput on one NVIDIA DGX Spark than established frameworks. The results point to hardware-aware local serving, but remain author-reported.","published_at":"2026-11-15","updated_at":"2026-11-15","record_date":"2026-11-15","date_kind":"published_at","topics":["inference"],"entities":["Hibana","NVIDIA","NVIDIA DGX Spark","vLLM","SGLang","TensorFold","Qwen","Nemotron"],"related_patterns":["harness-architecture-outlives-model-choice"],"novelty_label":"new_fact","what_changed":"Hibana introduces a distinct unified-memory inference engine for local agentic workloads, reporting up to 3.4× more tasks per hour and 3.9× higher decoding throughput on DGX Spark.","prior_refs":["https://newruntime.com/posts/nvidia-released-flare-a-local-inference-server-for-594e36/","https://newruntime.com/posts/run-ai-models-for-edge-application-development-with-3c1ef4/"],"source_url":"https://arxiv.org/abs/2611.10469","source_urls":["https://arxiv.org/abs/2611.10469"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:hibana-cornell-led-inference-engine-for-local-agentic-55553a","status":"published","visuals":[],"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/hibana-cornell-led-inference-engine-for-local-agentic-55553a/","markdown":"https://newruntime.com/posts/hibana-cornell-led-inference-engine-for-local-agentic-55553a.md","json":"https://newruntime.com/posts/hibana-cornell-led-inference-engine-for-local-agentic-55553a.json"},"markdown_url":"https://newruntime.com/posts/hibana-cornell-led-inference-engine-for-local-agentic-55553a.md","next_reads":[{"type":"topic","path":"/topics/inference/","url":"https://newruntime.com/topics/inference/","title":"Inference","reason":"Explore the inference topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/inference.md","json_url":"https://newruntime.com/topics/inference.json"},{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/sources/","url":"https://newruntime.com/sources/","title":"Sources","reason":"See the original sources this archive cites.","media_type":"text/html"}]}
