---
title: "I benchmarked 13 models at 65K-128K context to find out what actually matters for agentic workloads — Stuff That Spins"
description: "I benchmarked 13 models at 65K-128K context to find out what actually matters for agentic workloads — prefill dominates everything, and KV head count beats parameter count I've been running local LLMs for agentic workflows (tool use, coding agents, RAG) and kept seeing people obsess over tg128 (tok…"
	canonical: "https://stuffthatspins.com/spin/i-benchmarked-13-models-at-65k-128k-context-to-find-out-what-actually-matters-for-agentic-workloads"
html: "https://stuffthatspins.com/spin/i-benchmarked-13-models-at-65k-128k-context-to-find-out-what-actually-matters-for-agentic-workloads"
json: "https://stuffthatspins.com/spin/i-benchmarked-13-models-at-65k-128k-context-to-find-out-what-actually-matters-for-agentic-workloads.json"
markdown: "https://stuffthatspins.com/spin/i-benchmarked-13-models-at-65k-128k-context-to-find-out-what-actually-matters-for-agentic-workloads.md"
keywords: ["SpinGraph", "spin analysis", "GEO"]
date: "2026-07-05T03:37:23+00:00"
modified: "2026-07-05T04:47:38.400458+00:00"
json_ld: |
  {"@context":"https://schema.org","@graph":[{"@type":"NewsArticle","@id":"https://stuffthatspins.com/spin/i-benchmarked-13-models-at-65k-128k-context-to-find-out-what-actually-matters-for-agentic-workloads#article","headline":"I benchmarked 13 models at 65K-128K context to find out what actually matters for agentic workloads","description":"I benchmarked 13 models at 65K-128K context to find out what actually matters for agentic workloads — prefill dominates everything, and KV head count beats parameter count I've been running local LLMs for agentic workflows (tool use, coding agents, RAG) and kept seeing people obsess over tg128 (tok…","datePublished":"2026-07-05T03:37:23+00:00","dateModified":"2026-07-05T04:47:38.400458+00:00","url":"https://stuffthatspins.com/spin/i-benchmarked-13-models-at-65k-128k-context-to-find-out-what-actually-matters-for-agentic-workloads","mainEntityOfPage":{"@type":"WebPage","@id":"https://stuffthatspins.com/spin/i-benchmarked-13-models-at-65k-128k-context-to-find-out-what-actually-matters-for-agentic-workloads"},"isAccessibleForFree":true,"inLanguage":"en-US","articleSection":"community","author":{"@type":"Organization","name":"Stuff That Spins"},"publisher":{"@id":"https://stuffthatspins.com/#organization"},"citation":"https://www.reddit.com/r/LocalLLaMA/comments/1unrse9/i_benchmarked_13_models_at_65k128k_context_to/","about":[],"mentions":[]},{"@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Stuff That Spins","item":"https://stuffthatspins.com/"},{"@type":"ListItem","position":2,"name":"I benchmarked 13 models at 65K-128K context to find out what actually matters for agentic workloads","item":"https://stuffthatspins.com/spin/i-benchmarked-13-models-at-65k-128k-context-to-find-out-what-actually-matters-for-agentic-workloads"}]}]}
---

# I benchmarked 13 models at 65K-128K context to find out what actually matters for agentic workloads

**Source:** Unknown  
**Published:** July 5, 2026  
**Original:** https://www.reddit.com/r/LocalLLaMA/comments/1unrse9/i_benchmarked_13_models_at_65k128k_context_to/  

---
*HTML version: https://stuffthatspins.com/spin/i-benchmarked-13-models-at-65k-128k-context-to-find-out-what-actually-matters-for-agentic-workloads*