Local Qwen is useful, but not a cheap Opus replacement
Local Qwen 3.6 27B and Qwopus are not reliable replacements for Claude Opus or Codex on long, unsupervised coding work. In a small software business, they were useful for work that involved customer diagnostics and telemetry data that should not be sent to a cloud model. A 12000 dollar RTX 6000 Pro with 96GB of video memory paid for itself within a few months after local analysis found that one customer had under-reported license usage by about 4 to 5 times for more than a year.
The same setup still failed on broader agent tasks: it repeated outputs, invented file names and tool calls, made arithmetic mistakes, and sometimes drew the wrong business conclusion from usage data. Earlier 3090 cards forced heavier compression, shorter or lower-quality , and more fragile results. On the RTX 6000 Pro, Qwen 3.6 27B ran through llama.cpp with and high-quality settings, and MTP raised speed from about 67 to 130-200 in sustained use.
Running local AI became an operations problem, not just a model choice: the team needed , usage tracking, quotas, , power monitoring, and uptime. The practical lesson is to use local models for narrow support, maintenance, code reading, and testing tasks, while avoiding long unattended agent work.
Key points
- Local Qwen helped with private customer support and telemetry analysis, not as a full Opus or Codex replacement.
- A 12000 dollar RTX 6000 Pro paid for itself after finding a customer had under-reported license usage by 4 to 5 times.
- Local models were valuable when customer data could not reasonably be sent to a .
- Long unattended agent tasks caused loops, details, and unreliable conclusions.
- A real local AI setup needs , metering, quotas, routing, power tracking, and uptime planning.