mrkeyoor.com_
Sun 16 Aug 14:45 UTC
LLM Toolsevaluationupdated 16 Aug 2026

llmfit

llmfit is a terminal tool that detects your computer's memory, processor, and graphics hardware, then ranks hundreds of local language models by what should fit, how fast they may run, and what jobs they suit. It helps you avoid downloading a huge model only to discover that the chosen quantization or context size overwhelms your machine.

Verdict

llmfit is the first tool we would use to narrow a local-model shortlist, especially on unfamiliar hardware. Its transparent assumptions, quantization choices, provider integrations, and path from estimates to personal benchmarks make it much more useful than a static memory table. Treat its answer as a well-informed screening result, not a promise: verify available VRAM, serve the finalist, and measure your own prompt sizes before committing.

Setup5/5Many prebuilt install paths and useful results before model download
Docs5/5Clear formulas, platforms, providers, automation, and bench guides
Community5/5Large audience with rapid fixes and fresh benchmark submissions
Maturity4/5Polished release line, with hardware edge cases still surfacing

Who it’s for

Developers choosing a first local model for coding, reasoning, chat, vision, or embeddings.
Owners of NVIDIA, AMD, Intel, Apple Silicon, or CPU-only systems who want hardware-aware comparisons.
Ollama, llama.cpp, MLX, LM Studio, vLLM, or Docker Model Runner users who want discovery and measured benchmarks in one TUI.
Automation and agent builders who need JSON, REST, or MCP access to model-fit recommendations.

Who it’s NOT for

Buyers who need guaranteed throughput before acquiring hardware: unmeasured speed is a formula-based estimate, and the project recommends benchmarking a live provider for real numbers.
Operators evaluating a busy discrete GPU without checking it separately: an open issue shows GPU fit can use total VRAM rather than currently free VRAM, reporting a model as fitting on an occupied card.
Ryzen AI NPU or hybrid-runtime users: an open request confirms those paths and the Lemonade and FastFlowLM providers are not supported.
Android users expecting automatic accelerator detection: the platform guide says mobile GPUs are normally invisible and manual memory overrides only simulate recommendations.
People who want one binary to supply every inference runtime: llmfit can download and launch GGUF models through llama.cpp, but other paths still depend on installed and running providers.

Setup reality

Getting a recommendation is unusually easy because maintained installation paths include Scoop, Homebrew, MacPorts, a release script, Python tooling, containers, and Cargo. The model catalog is embedded, so the basic fit command does not require downloading a model. Accuracy still depends on hardware probes such as nvidia-smi, rocm-smi, sysfs, or system_profiler; manual overrides are available when detection is wrong. Downloads need a compatible provider or Hugging Face access, and real benchmarking requires a model already served by Ollama, vLLM, MLX, or llama.cpp. Sharing results is optional and invokes GitHub authentication to create a pull request.

A better first question than which model is best

Local model advice often begins with model quality and ends with an unpleasant memory error. llmfit reverses the order. It inspects RAM, CPU cores, GPU identity, VRAM, backend, and multi-GPU layout, then asks which model and quantization make sense on that actual machine. Results are ranked for general use, coding, reasoning, chat, multimodal work, or embeddings.

The default terminal interface shows fit, estimated speed, quality, and context together. Classic commands expose the same data as tables or JSON. You can inspect one model, simulate different hardware, cap the intended context, or ask what hardware a specific model requires. A REST service makes the analysis available to schedulers, while the source includes a standard-input MCP server and an OpenClaw advisor skill for agents. This breadth is useful because model choice is rarely only an interactive desktop question.

The estimates are unusually inspectable

The project does not present its ranking as magic. It embeds a catalog sourced from Hugging Face, identifies mixture-of-experts architectures, and tries quantizations from higher quality to more compressed until one fits. Memory scoring considers the model and context, with a second attempt at half context when necessary. Quality includes size, family reputation, quantization cost, use-case alignment, and curated family-level benchmark data.

Speed begins with a bandwidth model. For a recognized GPU, llmfit divides memory bandwidth by model size and applies an efficiency factor plus runtime adjustments. Its documentation names the default factor, explains the fallback constants for unknown hardware, and lets users tune assumptions. llmfit info exposes the inputs behind a number. That transparency is the project's strongest editorial feature: a prediction can be challenged instead of accepted as a mysterious score.

Predictions remain predictions. Prompt processing, cache length, runtime kernels, offload policy, model architecture, background load, and quantization implementation all affect observed performance. The right use is to eliminate obviously unsuitable downloads and produce a shortlist, not to promise an exact token rate.

It closes the loop with real measurements

The benchmarking workflow makes that limitation manageable. llmfit can discover models served through Ollama, llama.cpp, MLX, or vLLM, run three inference passes, and record token generation speed plus time to first token. Local results replace estimates on your machine and can calibrate other dense models on the same hardware.

Sharing is optional. When enabled, the tool uses GitHub's device flow or an environment token, forks the repository, writes result files, and opens or updates a pull request. Merged measurements ship with later releases, so an identical hardware profile can receive a community number before running its own test. The documented trust order puts personal results first, then matching community data, then external medians, then the formula. That hierarchy is sensible.

There is one operational distinction to understand. llmfit can download GGUF files and can launch an installed llama-cli or llama-server, but it is not a universal inference engine. Benchmarking requires a live provider, and Ollama, LM Studio, MLX, Docker Model Runner, and vLLM keep their own installation and serving responsibilities. Think of llmfit as the selection and verification layer across runtimes.

Hardware detection has real edge cases

Platform coverage is broad: Linux, Windows, Intel Mac, Apple Silicon, and CPU-only operation are documented. NVIDIA detection uses nvidia-smi, AMD uses rocm-smi, Intel relies on sysfs or lspci, and Apple unified memory comes from system_profiler. Overrides for VRAM, RAM, and core count let users model another machine or correct a failed probe.

The open issue tracker shows why verification matters. One current report demonstrates that the discrete-GPU path grades against total VRAM while CPU paths use available RAM. On the reporter's occupied 24 GB card, a model needing nearly the full card was labeled marginal even though roughly 1 GB was free. Until that is resolved, close GPU workloads or check free memory yourself before trusting a fit label.

Another open request covers Ryzen AI NPUs and hybrid runtimes that llmfit does not support. The platform guide also says Android GPU detection normally fails, with memory overrides serving only as a scoring simulation. Provider name mapping is maintained data, not a universal truth: an open Gemma issue shows an installed Ollama tag that does not map back to its catalog entry. None of these faults ruins the core workflow, but each marks a boundary around the automatic result.

Healthy, fast-moving, and worth installing

The repository was pushed on August 14, 2026, and version 1.1.9 was released on August 9. That release fixed Windows hardware detection, model ID mapping, mixture-of-experts speed estimation, plan consistency, and an unwanted dashboard side effect. Issue and pull-request activity continued through August 16, including hardware fixtures, community measurements, dependency work, and new model formats. The open count of 69 combines issues and pull requests.

Documentation is excellent. Separate guides explain the scoring formulas, platform boundaries, provider behavior, automation interface, custom models, and full benchmarking journey. More than 31,000 stars and nearly 2,000 forks are backed by visible maintenance rather than popularity alone.

Install llmfit before shopping for a local model or new GPU. Use it to compare plausible combinations, inspect why they ranked, and choose what to test. Then run the actual model at your intended context and concurrency. That final check is not a criticism of llmfit. It is exactly the workflow the project is designed to make cheaper.

Alternatives

ProjectWhat it isPick it when
llm-checkerA Node.js tool that pulls and benchmarks models through Ollama instead of relying mainly on estimates.pick this instead when you already use Ollama and prefer immediate measurements over broad MoE-aware catalog analysis.
OllamaA popular local model runtime with simple pull, run, and API workflows.pick this instead when you already know the model you want and need the runtime rather than a hardware advisor.
llama.cppA portable GGUF inference engine with detailed controls across CPUs and many GPU backends.pick this instead when low-level inference, conversion, or benchmarking is the job and model selection is secondary.
llmserveA companion terminal interface focused on selecting a backend and serving a chosen local model.pick this instead when you have picked a model and want a dedicated serving workflow.

What people are saying

  1. [github-trending] AlexsJones/llmfit

Sources

  1. llmfit README
  2. How llmfit works
  3. llmfit benchmarking guide
  4. llmfit v1.1.9 release
  5. Discrete GPU available-VRAM issue
  6. NPU and hybrid runtime request