skills.rest
tankpkg
tankpkgOFFICIAL
Explore all skills by this author →
GitHub

@tank/local-llm-fit

Determine if an open-source LLM fits a PC's VRAM, RAM, and bandwidth.

1|1|Updated Feb 20, 2026
Search Tags:#quantization#vram#local llm#kv cache#tokens per second#model fit#inference engines
One-click install
npx skills add https://github.com/tankpkg/packages --skill tank-local-llm-fit
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: @tank/local-llm-fit
Source: https://github.com/tankpkg/packages/tree/main/skills/local-llm-fit
Command: npx skills add https://github.com/tankpkg/packages --skill tank-local-llm-fit

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill decides whether a specific open-source LLM can run well on a given computer, replacing guesswork with clear fit, speed, and context-limit estimates.

Core Features & Use Cases

  • Exact model resolution: Searches for the exact model name the user gave and avoids silently swapping in a different release.
  • Fit and throughput analysis: Estimates VRAM and RAM usage, KV cache growth, and tokens per second from hardware bandwidth.
  • Engine and quantization guidance: Recommends the best inference engine and quantization format for the target machine.
  • Use case: A user asks whether a named model will run on an RTX 4090, an M2 Mac, or a CPU-only workstation, and this Skill returns a practical verdict with memory math.

Quick Start

Use the local-llm-fit skill to evaluate whether a specific model will run on my hardware and tell me the best quantization, engine, memory use, throughput, and verdict.

Frequently Asked Questions about @tank/local-llm-fit

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I check if a local LLM will fit in my VRAM and RAM?▼

Calculating local LLM fit involves analyzing exact model weights, KV cache growth, and memory bandwidth to estimate VRAM and RAM usage. This determines whether your hardware can handle the inference workload efficiently.

How do I calculate tokens per second for local inference engines?▼

Calculate tokens per second for local inference engines by evaluating hardware memory bandwidth against the model memory footprint. This estimates actual inference throughput for your specific desktop, laptop, or GPU rig.

Can I run an open-source LLM on a Mac or CPU-only workstation?▼

Yes, you can run an open-source LLM on a Mac or CPU-only workstation by matching the right quantization format and inference engine. Fit analysis evaluates hardware specs to provide a practical compatibility and speed verdict.

What is the best quantization format for running local LLMs on limited hardware?▼

The best quantization format for running local LLMs on limited hardware depends on your exact model resolution and target machine. Fit analysis recommends the optimal quantization format and engine to balance memory usage and inference speed.

How does KV cache growth affect long-context local LLM workloads?▼

KV cache growth affects long-context local LLM workloads by increasing VRAM and RAM requirements as the context window expands. Estimating this memory growth is essential to determine if hardware can sustain long-context inference.

Related Skills

Explore More Software Engineering Skills→
CxxxxDxxxFCxxxxDxxxF

llmfit-advisor

Analyze local hardware to recommend LLMs with quantization and speed estimates.

Community
Intermediate
peteromalletpeteromallet
97

serving-llms-vllm

Deploy LLM inference with vLLM using OpenAI-compatible endpoints, quantization, and tensor parallelism.

Community
Intermediate
ovachieverovachiever
52

tensorrt-llm

Optimize LLM inference on NVIDIA GPUs with TensorRT-LLM.

Community
Advanced
GeargrindadminGeargrindadmin

llmfit-advisor

Analyze local CPU, RAM, and GPU/VRAM to recommend LLM models with quantization and fit scores.

Community
Intermediate