{"type":"post","id":"macseek-from-scratch-metal-engine-runs-the-233-005b3d","slug":"macseek-from-scratch-metal-engine-runs-the-233-005b3d","title":"MacSeek: from-scratch Metal engine runs the 233 GB DeepSeek-V4.1-Flash GGUF with 200K context on a 64 GB Mac","description":"MacSeek is an independent developer’s from-scratch Metal inference engine for running DeepSeek-V4.1-Flash from a 233 GB quantized GGUF on a 64 GB Apple Silicon Mac.","retrieval_nugget":"MacSeek is an independent developer’s from-scratch Metal inference engine for running DeepSeek-V4.1-Flash from a 233 GB quantized GGUF on a 64 GB Apple Silicon Mac. Its README reports substantially faster long-context prefill and decode than llama.cpp on one M5 Pro, making specialized SSD streaming a concrete local-inference approach worth testing—but not yet an independently verified result.","published_at":"2026-10-31","updated_at":"2026-10-31","record_date":"2026-10-31","date_kind":"published_at","topics":["harness-engineering","agent-runtime"],"entities":["MacSeek","DeepSeek-V4.1-Flash","Arnav Prabhu","llama.cpp","Unsloth"],"related_patterns":["harness-architecture-outlives-model-choice"],"novelty_label":"new_fact","what_changed":"MacSeek adds a Mac-focused, from-scratch Metal engine that runs the 233 GB model with 200K context on 64 GB RAM, claiming 7× faster prefill and 12× faster decode than llama.cpp.","prior_refs":["memory:1464f1a1-c7ff-89b9-aefd-f8f70bf8ed7c","memory:cc12dad7-b7d8-85bd-a3ad-1ab9097d35b6"],"source_url":"https://github.com/arnavprabhu/MacSeek","source_urls":["https://github.com/arnavprabhu/MacSeek"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:macseek-from-scratch-metal-engine-runs-the-233-005b3d","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/macseek-from-scratch-metal-engine-runs-the-233-005b3d/","markdown":"https://newruntime.com/posts/macseek-from-scratch-metal-engine-runs-the-233-005b3d.md","json":"https://newruntime.com/posts/macseek-from-scratch-metal-engine-runs-the-233-005b3d.json"},"next_reads":[{"type":"topic","path":"/topics/harness-engineering/","url":"https://newruntime.com/topics/harness-engineering/","title":"Harness Engineering","reason":"Explore the harness-engineering topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/harness-engineering.md","json_url":"https://newruntime.com/topics/harness-engineering.json"},{"type":"topic","path":"/topics/agent-runtime/","url":"https://newruntime.com/topics/agent-runtime/","title":"Agent Runtime","reason":"Explore the agent-runtime topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/agent-runtime.md","json_url":"https://newruntime.com/topics/agent-runtime.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":"/source-ledger.json","url":"https://newruntime.com/source-ledger.json","title":"Source Ledger","reason":"Inspect the public source evidence graph.","media_type":"application/json"}]}