AI Model Guides

Data-driven guides for picking the right AI model. Each one is generated from real benchmarks, file sizes, and download counts — not opinion. Re-renders whenever the underlying data refreshes.

Best LLMs Under 8GB VRAM

Top language models that fit on entry-level GPUs and laptops with 8GB.

Best LLMs for 12GB VRAM

The sweet spot — 12GB cards run most 7–13B models comfortably.

Best LLMs for 24GB VRAM

What the RTX 3090, 4090, and 7900 XTX can run.

Best Coding Models You Can Run Locally

Top open-source code models for completion, refactoring, and explanation.

Best Open-Source Image Generation Models

Stable Diffusion, Flux, and the new wave of open image models.

Best Tiny Models (Under 2B Params)

Models small enough to run on phones, browsers, and edge devices.

Best Vision-Language Models

Multimodal models that read images, screenshots, and diagrams.

Best Speech Recognition (Whisper) Models

Open-source speech-to-text — from Whisper Tiny to Large v3 Turbo.

Best Open-Source Text-to-Speech Models

From XTTS v2 to Kokoro — local TTS that sounds human.

Best Embedding Models for RAG

Embedding and reranker models for retrieval-augmented generation.

Best Uncensored / Abliterated Models

Models with safety alignment removed for unrestricted output.

Best Long-Context Models (32K+)

Models that handle whole books, codebases, and meeting transcripts.

Best Models for the RTX 4090

What 24GB on Ada Lovelace actually runs — and how fast.

Best Models for the RTX 4060

Best AI workloads for an entry-level Ada Lovelace card.

Best Models for Apple Silicon

MLX-friendly picks for M-series Macs.