DEDUPEDWho it is forA hobbyist comparing models on a home GPU.
The problemPublished speed numbers rarely match your hardware.
The outcomeMeasured tok/s saved locally that replace estimates.
What you need first
- macOS, Windows, Linux or Docker
Steps
- Open the llmfit TUI
- Choose a model to download and serve
- Run the benchmark
- Optionally contribute results as a PR from the TUI
Prompt for your agent
Use llmfit to benchmark <model> on this machine and report measured tokens per second versus the estimate.
Caveat. Benchmarks download full models, which can be many gigabytes.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for AlexsJones/llmfit and "Measure real tokens per second on your hardware"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer