Stay informed with weekly updates on the latest AI tools. Get the newest insights, features, and offerings right in your inbox!
What LLM Can I Run helps you find and rank large language models by VRAM fit and real benchmark scores for any GPU.
What LLM Can I Run takes the guesswork out of running large language models locally. Tell it your GPU — RTX 4090, M3 Max 64GB, A6000, anything — and it ranks every model that actually fits in your VRAM, scored by real benchmarks (LiveBench, Aider Polyglot, Chatbot Arena ELO). No vague claims, no signup, no fluff. See sweet-spot quants (Q4_K_M vs Q6_K vs FP16), tok/s estimates, fit grades (great / ok / tight), and the upgrade path that unlocks the next tier. Apple Silicon's unified-memory math is accounted for. NVIDIA, AMD, and Apple Silicon supported out of the box.